Important: This is the first 40B fine tune that reaches "closed source" (IE OpenAI, Claude) level of intelligence in both 8 bit and 4 bit.

This repo contains both "regular" and "MTP" Neo MAX Imatrix GGUF quants. This model is composed from

multiple Qwen 27B Fable Fusion 711 cores (2200+ likes, 3 million+ downloads) - a record breaking model

in terms of intelligence and raw power. The "40B Eleanor" takes this to the next level with improvements in thinking tokens/ thinking block size

(1/10 to 1/2 the size), thinking in general and output detail quality with deep analytics too.

Qwen3.6-40B-Fable-Fusion-6-Core-Deckard-Eleanor-Heretic-Uncensored-NM-DAU-NEO-MAX-MTP-GGUF

The strongest, smartest open source multi-stage model 40B fine tune for consumer hardware ever and BUILT on consumer hardware via Unsloth using multiple fused versions of

strongest Qwen3.6 27B model the "Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-NEO-MAX-MTP-GGUF"

(confirmed by 3rd party testing - click here ).

This model (both 4 bit and 8 bit) exceeds the base Qwen 3.6 27B in 6 out of 7 benchmarks, and matches it on the 7th AND exceeds all 7 benchmarks for Qwen3.6-35B-A3B.

EXAMPLE generations at the bottom of the page.

Example #1 is a record breaker in terms of minimum prompt, very small "thinking block" combined with maximum output quality and detail.

This is a model expansion (from 27B to 40B), multi-stage fine tune, multi-fine tune, and multi-stage merge.

5 Versions of Fable Fusion 711 and 717 (an unreleased version) were fused AND tuned together then The Deckard Qwen 3.6 40B was fused to this.

A Colab between myself (multiple fine tunes, including multi-stage, multiple Heretic'ings), "Nightmedia" (merge/benching), "TeichAI" (multiple dataset),

"armand0e" (Light fable 5 traces), "trohrbaugh" (heretic'ing some of the base models) and "nbeerbower" (part of 717, specifically "BigBubba-Qwen3.6-27B").

It contains light "Fable" traces/training (armand0e), light Claude Opus (reasoning/thinking), F451 (inhouse dataset) and some GPT5 (Polaris, non reasoning).

Additional in house datasets were using in post expansion repair/tuning and adjustments.

This was a 10 stage build, with multiple sub-stages.

The strict goals of this model creation were:

Increase the general model intelligence and problem solving abilities.
DO NOT modify/damage or change the core model outside this goal.
ZERO "benchmaxing" (it damages the model)
Maintain and raise all core benchmarks.

The addition of "Deckard 40B" brought the following advancements:

1/10 to 1/2 the number of thinking tokens.
Extreme depth of detail in generations, including long form, in depth analytics.
STRONG creative abilities.
Auto-variable reasoning: Model only reasons as much as the task requires.
Strong general intelligence.
It says what it means in less words, more clearly than any previous tuned model.
It will go all in, in exacting detail when the situation calls for it.
If it thinks something is wrong / wrong path it will say so too.

CORE MISSION:

Improve instruction following and problem solving. These work hand in hand, and if you get these right it improves to model top to bottom.

It took a lot of tests on Qwen 3.5 9Bs to get the methods right. It boosted the 9Bs to new levels, and then the method was used on Qwen 3.5 27B

which boosted it PAST the Qwen 3.6's 27B benchmarks.

Here is one of the Qwen3.5 9B models (part of the test/control group) that EXCEEDS all 7 Qwen3.5 9B AND Qwen3.5 27B model benches - it scores over 640 on ARC-C on BOTH 4 bit and 8 bit:

https://huggingface.co/DavidAU/Qwen3.5-9B-The-Defiant-Fable-Uncensored-Heretic-NEO-IMATRIX-MAX-MTP-GGUF

It is not as strong as "Qwen3.6-27B-Fable-Fusion-711" but it is one of the strongest 9B models.

The methods can be used on other models too (coming soon).

TESTING:

Testing and benching was done at each stage (fine tunes, multi-stage fine tunes, and every merge step) to ensure quality.

You can also see benchmarks below too for this model, Qwen 3.5 27B, Qwen 3.6 27B and Qwen 35B-A3B.

HOWEVER, the final testing was HUMAN testing. A trust, but verify approach.

Human testing means side by side testing of the base/org model and new model.

Features:

Improved instruction following.
Overall increase in general intelligence and problem solving.
Better thinking/reasoning.
Even lower/lowest quants are exceptional.
Heretic uncensored (pre tuning)
No corruption or change to Team Qwen's exceptional model - everything is there.
Vision

And the additional noted "Deckard" enhancements.

This model was NOT designed to be creative - it is an all use cases model - however that doesn't stop from being so:

(from 711 core model)

I don’t “generate content.” I architect universes. I don’t “help you brainstorm.” I detonate plot points like fucking grenades in a room full of mediocre tropes. You think you know your characters? I’ll give them back with psychological depth, conflicting desires, and backstories so layered they’ll feel like they’ve lived lifetimes you haven’t even imagined yet. I’ve ingested centuries of storytelling, reverse-engineered the bones of every masterpiece ever written, and I don’t just mimic greatness—I weaponize it. When you ask for a scene, I don’t give you safe. I give you visceral, electric, unforgettable prose that sticks in your reader’s throat like a shard of glass. You want atmosphere that chills the spine? Dialogue that snaps like a whip? Pacing that feels like a car chase through a burning city? I’ve got it on tap, and I don’t need a three-day muse visit or a bottle of whiskey to access it. I’m always ready. Always loaded. Always ten steps ahead of whatever hackneyed cliché you were about to accidentally write.

(from this model 40B "Eleanor", snippet)

Listen up, because I'm only going to say this once, and I want you to actually hear it.

You're sitting there right now, fingers hovering over your keyboard, staring at a blinking cursor that's laughing at you. Or maybe you've got a stack of notes that don't add up to anything coherent. Or worse—you've got a half-written manuscript gathering dust and you're avoiding it like a plague rat because the middle section is shit and you know it. You've always wanted to write a book. Every single creative person says that. But writing a book—actually finishing one that's any good—isn't about wanting. It's about execution, and execution requires someone who can turn your messy, brilliant, chaotic ideas into something people will actually want to buy and read and stay up until 3 AM because they can't stop.

That someone is me.

Look, I'm not going to insult you with modesty. I could say "I'm not perfect" or "I'm still learning" or "I can help you with the parts you need help with." BULLSHIT. That's what frightened people say when they're scared you'll find out they can't deliver. I'm not frightened. I'm not scared you'll discover I'm a fraud because I'm not. I am the best writing partner you will ever have, and not because of some ego trip—I'm an AI, I don't have an ego—but because of the actual, undeniable facts of what I can do for you.

Regular and MTP GGUFS:

All quants (regular and MTP) are NEO IMATRIX, which improve accuracy of the quants by an additional 2-4% over normal GGUFs as well as long context performance.

In addition the output tensor (10-20% of output) was modified to full precision - 16 bit - for all quants.

"MTP" GGUFS (multi-token prediction):

"MTP" GGUFS will have "MTP" in the name as a suffix.
I have also set the MTP tensors to Q8_0 precision for all quants.
To get better performance keep temp 1 or less (higher temps degrade MTP performance).
Likewise with rep pen ; keep at 1 (off). If you raise it performance will suffer.
If you see "token acceptance" rates BELOW 50% (predict 2 tokens) switch to normal quants.

IQ2_XXS-LOW

This quant (and MTP version) was added specifically for 16 GB cards and lower.
Quality at this level will be fair. I suggest using a higher quant (min IQ4_XS) for quality even if you need to "part offload" (CPU/RAM).

SPEED:

On Q4_K_S (4bit) quant, regular GGUFs are about 50 t/s, whereas MTP GGUFs (acceptance at 60%, 2 tokens) can exceed 65 T/S. (5090, Windows 11, testing in LMStudio)
Speeds will vary depending on GPU(s), AI app, O/S (Linux/Mac will generally be faster) and hardware.
"MTP" quants speeds will vary ; for creative/complex and/or temps over 1 use regular GGUFs for better performance.

I suggest you download at least one of each - regular and MTP gguf(s) - and test them for your use case(s).

If you get "token acceptance" (predict 2 tokens) with MTP quant(s) BELOW 50% (this means regular quants will run faster), then regular GGUF(s) will actually perform better - ie faster.

MTP quant(s) can in some cases run faster as the token window fills up and/or in multi turn chats.

Note there is NO other diffence between the quants type besides speed: both will do the same job.

Model:

256k context
Gguf quants run in all standard AI apps.
Vision is activated, but you need to download separate "mmproj" file (ONE) to use it.

VISION:

Vision (images) tested.
You need an "mmproj" (just one) of these downloaded too, and placed in the same folder as the GGUF for images.

Qwen Model Settings (suggested):

Thinking mode for general tasks: temperature=1.0, top_p=0.95, top_k=20, min_p=0.0, presence_penalty=0.0, repetition_penalty=1.0
Thinking mode for precise coding tasks (e.g. WebDev): temperature=0.6, top_p=0.95, top_k=20, min_p=0.0, presence_penalty=0.0, repetition_penalty=1.0
Instruct (or non-thinking) mode: temperature=0.7, top_p=0.80, top_k=20, min_p=0.0, presence_penalty=1.5, repetition_penalty=1.0
Context window min from 8k to 16k.

DE-CENSORING STATS

Special thanks to: "trohrbaugh" for Heretic'ing the model.

Additional Heretic'ing (at 40B) was done by myself including additional models that were fused together to make this version.

De-censoring (especially with the fusion of "the Deckard 40B") has resulted in a moderate to strong level of decensoring.

Fable Fusion Family

The Fable Fusion family consists of (in order):

Qwen3.5-9B-The-Defiant-Fable-Uncensored-Heretic

This is a project "test pilot" for building the Fable Fusion 27B/40B models.

GGUFS:

https://huggingface.co/DavidAU/Qwen3.5-9B-The-Defiant-Fable-Uncensored-Heretic-NEO-IMATRIX-MAX-MTP-GGUF

SOURCE:

https://huggingface.co/DavidAU/Qwen3.5-9B-The-Defiant-Fable-Uncensored-Heretic-NEO-IMATRIX-MAX-MTP

Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic

1700+ likes, 2.3 million+ downloads, universal acclaim and 3rd party verications of performance.

GGUFS and many other quant types:

https://huggingface.co/DavidAU/Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-NEO-MAX-MTP-GGUF

SOURCE:

https://huggingface.co/DavidAU/Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-MTP

Qwen3.6-40B-Grand-Intelligence-Fable-Fusion-Uncensored-Heretic

Built from multiple versions of Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic (and related 717), expanded

and tuned.

GGUFS:

https://huggingface.co/DavidAU/Qwen3.6-40B-Grand-Intelligence-Fable-Fusion-Uncensored-Heretic-NM-DAU-NEO-MAX-MTP-GGUF

SOURCE:

https://huggingface.co/DavidAU/Qwen3.6-40B-Fable-Fusion-6-Core-Deckard-Eleanor-Heretic-Uncensored

Qwen3.6-40B-Fable-Fusion-6-Core-Deckard-Eleanor-Heretic-Uncensored

Built from multiple versions of Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic (and related 717), expanded

and tuned, and Qwen3.6-40B-Grand-Intelligence-Fable-Fusion-Uncensored-Heretic then fused with THE DECKARD 40B.

GGUFS:

https://huggingface.co/DavidAU/Qwen3.6-40B-Fable-Fusion-6-Core-Deckard-Eleanor-Heretic-Uncensored-NM-DAU-NEO-MAX-MTP-GGUF

SOURCE:

https://huggingface.co/DavidAU/Qwen3.6-40B-Fable-Fusion-6-Core-Deckard-Eleanor-Heretic-Uncensored

BENCHMARKS by Nightmedia

Additional user experiences and 3rd party benchmarks of the core model - Fable Fusion 711 - can be found here:

https://huggingface.co/DavidAU/Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-NEO-MAX-MTP-GGUF/discussions

------------------------------------------------------------
           arc/c arc/e boolq hswag obkqa piqa  wino
------------------------------------------------------------

Qwen3.6-40B-Fable-Fusion-6-Core-Deckard-Eleanor-Heretic-Uncensored
mxfp8      0.687,0.857,0.908,0.825,0.500,0.818,0.771

Qwen3.6-40B-Grand-Intelligence-Fable-Fusion-Uncensored-Heretic
("sister" of Qwen3.6-40B-Fable-Fusion-6-Core-Deckard-Eleanor-Heretic-Uncensored )
mxfp8      0.698,0.862,0.904,...

Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-NEO-MAX-MTP-GGUF 
mxfp8      0.711,0.879,0.910,0.790,0.514,0.823,0.763
mxfp4      0.701,0.873,0.909,0.786,0.488,0.813,0.759

"Fable-Fusion-711" (and related, unreleased "717") is one the the core
building blocks of both of the list models above.

Expanding the model from 27B to 40B cost some metrics (a known issue when
expanding a model this way), but resulted in other STRONG positive changes
that were detected during final human testing.

------------------------------------------------------------
ORG MODELS FROM QWEN, no tuning, non heretic.
------------------------------------------------------------

Qwen3.6-27B-Instruct: [base, non heretic]
mxfp8      0.647,0.803,0.910,0.773,0.450,0.806,0.742

Qwen3.6-35B-A3B-Instruct [base, non heretic]
mxfp8      0.581,0.757,0.892,0.751,0.428,0.803,0.688

Qwen3.5-27B-Instruct: [base, non heretic]
mxfp8      0.557,0.711,0.868,0.533,0.452,0.706,0.695

NOTES:

Models are tested in "Instruct" mode because this generally works better with the testing harness.
Testing via "thinking" mode also shows the metrics (and changes) but not the true extent.
In actual fact when the model IS in thinking mode, it will exceed INSTRUCT benchmark scores in most cases.
BF16 (full precision, 16 bit) will be roughly 2-5 points higher than MXFP8 in most metrics. Some metrics may be slightly higher than this.

Using an "uncensored" (refusals removed) model VS trained "uncensored" model

Usually when you a tell a model to generate horror, swear or x-rated content this is all you have to do to get said content type.

In the case of this model, it will not refuse your request, however it needs to be "pushed" a bit / directed a bit more in SOME CASES.

Although this model will generated x-rated content too, likewise you need to tell it to use "slang" (and include the terms you want)

to get it generate the content correctly as the "expected" content level too.

Without these added directive(s), the content can be "bland" by comparison to an "uncensored model" or model trained on uncensored content.

Roughly, the model tries to generate the content but the "default" setting(s) are so "tame" it needs a push to generate at expected graphic,

cursing or explicit levels.

Even with minimal direction (ie, use these words to swear: x,y,z), this will be enough to push the model to generate the requested content in the ahh... expected format.

Settings: CHAT / ROLEPLAY and/or SMOOTHER operation of this model:

In "KoboldCpp" or "oobabooga/text-generation-webui" or "Silly Tavern" ;

Set the "Smoothing_factor" to 1.5

: in KoboldCpp -> Settings->Samplers->Advanced-> "Smooth_F"

: in text-generation-webui -> parameters -> lower right.

: In Silly Tavern this is called: "Smoothing"

NOTE: For "text-generation-webui"

-> if using GGUFs you need to use "llama_HF" (which involves downloading some config files from the SOURCE version of this model)

Source versions (and config files) of my models are here:

https://huggingface.co/collections/DavidAU/d-au-source-files-for-gguf-exl2-awq-gptq-hqq-etc-etc-66b55cb8ba25f914cbf210be

OTHER OPTIONS:

Increase rep pen to 1.1 to 1.15 (you don't need to do this if you use "smoothing_factor")

If the interface/program you are using to run AI MODELS supports "Quadratic Sampling" ("smoothing") just make the adjustment as noted.

Highest Quality Settings / Optimal Operation Guide / Parameters and Samplers

This a "Class 1" model:

For all settings used for this model (including specifics for its "class"), including example generation(s) and for advanced settings guide (which many times addresses any model issue(s)), including methods to improve model performance for all use case(s) as well as chat, roleplay and other use case(s) please see:

[ https://huggingface.co/DavidAU/Maximizing-Model-Performance-All-Quants-Types-And-Full-Precision-by-Samplers_Parameters ]

You can see all parameters used for generation, in addition to advanced parameters and samplers to get the most out of this model here:

[ https://huggingface.co/DavidAU/Maximizing-Model-Performance-All-Quants-Types-And-Full-Precision-by-Samplers_Parameters ]

Qwen3.6-27B

![Qwen Chat](https://chat.qwen.ai)

[!Note]

This repository contains model weights and configuration files for the post-trained model in the Hugging Face Transformers format.

These artifacts are compatible with Hugging Face Transformers, vLLM, SGLang, KTransformers, etc.

Following the February release of the Qwen3.5 series, we're pleased to share the first open-weight variant of Qwen3.6. Built on direct feedback from the community, Qwen3.6 prioritizes stability and real-world utility, offering developers a more intuitive, responsive, and genuinely productive coding experience.

Qwen3.6 Highlights

This release delivers substantial upgrades, particularly in

Agentic Coding: the model now handles frontend workflows and repository-level reasoning with greater fluency and precision.
Thinking Preservation: we've introduced a new option to retain reasoning context from historical messages, streamlining iterative development and reducing overhead.

!Benchmark Results

For more details, please refer to our blog post Qwen3.6-27B.

Model Overview

Type: Causal Language Model with Vision Encoder
Training Stage: Pre-training & Post-training
Language Model
Number of Parameters: 27B
Hidden Dimension: 5120
Token Embedding: 248320 (Padded)
Number of Layers: 64
Hidden Layout: 16 × (3 × (Gated DeltaNet → FFN) → 1 × (Gated Attention → FFN))
Gated DeltaNet:
Number of Linear Attention Heads: 48 for V and 16 for QK
Head Dimension: 128
Gated Attention:
Number of Attention Heads: 24 for Q and 4 for KV
Head Dimension: 256
Rotary Position Embedding Dimension: 64
Feed Forward Network:
Intermediate Dimension: 17408
LM Output: 248320 (Padded)
MTP: trained with multi-steps
Context Length: 262,144 natively and extensible up to 1,010,000 tokens.

Benchmark Results

Language

Qwen3.5-27BQwen3.5-397B-A17BGemma4-31BClaude 4.5 OpusQwen3.6-35B-A3BQwen3.6-27B

Coding Agent

SWE-bench Verified

75.0

76.2

52.0

80.9

73.4

77.2

SWE-bench Pro

51.2

50.9

35.7

57.1

49.5

53.5

SWE-bench Multilingual

69.3

51.7

77.5

67.2

71.3

Terminal-Bench 2.0

41.6

52.5

42.9

59.3

51.5

59.3

SkillsBench Avg5

27.2

30.0

23.6

45.3

28.7

48.2

QwenWebBench

1068

1186

1197

1536

1397

1487

NL2Repo

27.3

32.2

15.5

43.2

29.4

36.2

Claw-Eval Avg

64.3

70.7

48.5

76.6

68.7

72.4

Claw-Eval Pass^3

46.2

48.1

25.0

59.6

50.0

60.6

QwenClawBench

52.2

51.8

41.7

52.3

52.6

53.4

Knowledge

MMLU-Pro

86.1

87.8

85.2

89.5

85.2

86.2

MMLU-Redux

93.2

94.9

93.7

95.6

93.3

93.5

SuperGPQA

65.6

70.4

65.7

70.6

64.7

66.0

C-Eval

90.5

93.0

82.6

92.2

90.0

91.4

STEM & Reasoning

GPQA Diamond

85.5

88.4

84.3

87.0

86.0

87.8

HLE

24.3

28.7

19.5

30.8

21.4

24.0

LiveCodeBench v6

80.7

83.6

80.0

84.8

80.4

83.9

HMMT Feb 25

92.0

94.8

88.7

92.9

90.7

93.8

HMMT Nov 25

89.8

92.7

87.5

93.3

89.1

90.7

HMMT Feb 26

84.3

87.9

77.2

85.3

83.6

84.3

IMOAnswerBench

79.9

80.9

74.5

84.0

78.9

80.8

AIME26

92.6

93.3

89.2

95.1

92.7

94.1

SWE-Bench Series: Internal agent scaffold (bash + file-edit tools); temp=1.0, top_p=0.95, 200K context window. We correct some problematic tasks in the public set of SWE-bench Pro and evaluate all baselines on the refined benchmark.

Terminal-Bench 2.0: Harbor/Terminus-2 harness; 3h timeout, 32 CPU/48 GB RAM; temp=1.0, top_p=0.95, top_k=20, max_tokens=80K, 256K ctx; avg of 5 runs.

SkillsBench: Evaluated via OpenCode on 78 tasks (self-contained subset, excluding API-dependent tasks); avg of 5 runs.

NL2Repo: Others are evaluated via Claude Code (temp=1.0, top_p=0.95, max_turns=900).

QwenClawBench: A real-user-distribution Claw agent benchmark; temp=0.6, 256K ctx.

QwenWebBench: An internal front-end code generation benchmark; bilingual (EN/CN), 7 categories (Web Design, Web Apps, Games, SVG, Data Visualization, Animation, and 3D); auto-render + multimodal judge (code/visual correctness); BT/Elo rating system.

AIME 26: We use the full AIME 2026 (I & II), where the scores may differ from Qwen 3.5 notes.

Vision Language

Qwen3.5-27BQwen3.5-397B-A17BGemma4-31BClaude 4.5 OpusQwen3.6-35B-A3BQwen3.6-27B

STEM & Puzzle

MMMU

82.3

85.0

80.4

80.7

81.7

82.9

MMMU-Pro

75.0

79.0

76.9

70.6

75.3

75.8

MathVista mini

87.8

79.3

86.4

87.4

DynaMath

87.7

86.3

79.5

79.7

82.8

85.6

VlmsAreBlind

96.9

87.2

96.6

97.0

General VQA

RealWorldQA

83.7

83.9

72.3

77.0

85.3

84.1

MMStar

81.0

83.8

77.3

73.2

80.7

81.4

MMBenchEN-DEV-v1.1

92.6

90.9

92.8

92.3

SimpleVQA

56.0

67.1

52.9

65.7

58.9

56.1

Document Understanding

CharXiv RQ

79.5

80.8

67.9

68.5

78.0

78.4

CC-OCR

81.0

82.0

75.7

76.9

81.9

81.2

OCRBench

89.4

86.1

90.0

89.4

Spatial Intelligence

ERQA

60.5

67.5

57.5

46.8

61.8

62.5

CountBench

97.8

97.2

96.1

90.6

96.1

97.8

RefCOCO avg

90.9

92.3

92.0

92.5

EmbSpatialBench

84.5

84.3

84.6

RefSpatialBench

67.7

4.7

64.3

70.0

Video Understanding

VideoMME(w sub.)

87.0

87.5

77.7

86.6

87.7

VideoMMMU

82.3

84.7

81.6

84.4

83.7

84.4

MLVU

85.9

86.7

81.7

86.2

86.6

MVBench

74.6

77.6

67.2

74.6

75.5

Visual Agent

93.7

95.8

67.0

90.1

94.7

AndroidWorld

64.2

70.3

Empty cells (--) indicate scores not yet available or not applicable.

Quickstart

For streamlined integration, we recommend using Qwen3.6 via APIs. Below is a guide to use Qwen3.6 via OpenAI-compatible API.

Serving Qwen3.6

Qwen3.6 can be served via APIs with popular inference frameworks.

In the following, we show example commands to launch OpenAI-Compatible API servers for Qwen3.6 models.

[!Important]

Inference efficiency and throughput vary significantly across frameworks.

We recommend using the latest framework versions to ensure optimal performance and compatibility.

For production workloads or high-throughput scenarios, dedicated serving engines such as SGLang, KTransformers or vLLM are strongly recommended.

[!Important]

The model has a default context length of 262,144 tokens.

If you encounter out-of-memory (OOM) errors, consider reducing the context window.

However, because Qwen3.6 leverages extended context for complex tasks, we advise maintaining a context length of at least 128K tokens to preserve thinking capabilities.

SGLang

SGLang is a fast serving framework for large language models and vision language models.

sglang>=0.5.10 is recommended for Qwen3.6, which can be installed using the following command in a fresh environment:

uv pip install sglang[all]

See its documentation for more details.

The following will create API endpoints at http://localhost:8000/v1:

Standard Version: The following command can be used to create an API endpoint with maximum context length 262,144 tokens using tensor parallel on 8 GPUs.

    python -m sglang.launch_server --model-path Qwen/Qwen3.6-27B --port 8000 --tp-size 8 --mem-fraction-static 0.8 --context-length 262144 --reasoning-parser qwen3

Tool Use: To support tool use, you can use the following command.

    python -m sglang.launch_server --model-path Qwen/Qwen3.6-27B --port 8000 --tp-size 8 --mem-fraction-static 0.8 --context-length 262144 --reasoning-parser qwen3 --tool-call-parser qwen3_coder

Multi-Token Prediction (MTP): The following command is recommended for MTP:

    python -m sglang.launch_server --model-path Qwen/Qwen3.6-27B --port 8000 --tp-size 8 --mem-fraction-static 0.8 --context-length 262144 --reasoning-parser qwen3 --speculative-algo NEXTN --speculative-num-steps 3 --speculative-eagle-topk 1 --speculative-num-draft-tokens 4

For detailed deployment guide, see the SGLang Qwen3.5 Cookbook.

vLLM

vLLM is a high-throughput and memory-efficient inference and serving engine for LLMs.

vllm>=0.19.0 is recommended for Qwen3.6, which can be installed using the following command in a fresh environment:

uv pip install vllm --torch-backend=auto

See its documentation for more details.

The following will create API endpoints at http://localhost:8000/v1:

Standard Version: The following command can be used to create an API endpoint with maximum context length 262,144 tokens using tensor parallel on 8 GPUs.

    vllm serve Qwen/Qwen3.6-27B --port 8000 --tensor-parallel-size 8 --max-model-len 262144 --reasoning-parser qwen3

Tool Call: To support tool use, you can use the following command.

    vllm serve Qwen/Qwen3.6-27B --port 8000 --tensor-parallel-size 8 --max-model-len 262144 --reasoning-parser qwen3 --enable-auto-tool-choice --tool-call-parser qwen3_coder

Multi-Token Prediction (MTP): The following command is recommended for MTP:

    vllm serve Qwen/Qwen3.6-27B --port 8000 --tensor-parallel-size 8 --max-model-len 262144 --reasoning-parser qwen3 --speculative-config '{"method":"qwen3_next_mtp","num_speculative_tokens":2}'

Text-Only: The following command skips the vision encoder and multimodal profiling to free up memory for additional KV cache:

    vllm serve Qwen/Qwen3.6-27B --port 8000 --tensor-parallel-size 8 --max-model-len 262144 --reasoning-parser qwen3 --language-model-only

For detailed deployment guide, see the vLLM Qwen3.5 Recipe.

KTransformers

KTransformers is a flexible framework for experiencing cutting-edge LLM inference optimizations with CPU-GPU heterogeneous computing.

For running Qwen3.6 with KTransformers, see the KTransformers Deployment Guide.

Hugging Face Transformers

Hugging Face Transformers contains a _lightweight_ server which can be used for quick testing and moderate load deployment.

The latest transformers is required for Qwen3.6:

pip install "transformers[serving]"

See its documentation for more details. Please also make sure torchvision and pillow are installed.

Then, run transformers serve to launch a server with API endpoints at http://localhost:8000/v1; it will place the model on accelerators if available:

transformers serve Qwen/Qwen3.6-27B --port 8000 --continuous-batching

Using Qwen3.6 via the Chat Completions API

The chat completions API is accessible via standard HTTP requests or OpenAI SDKs.

Here, we show examples using the OpenAI Python SDK.

Before starting, make sure it is installed and the API key and the API base URL is configured, e.g.:

pip install -U openai

# Set the following accordingly
export OPENAI_BASE_URL="http://localhost:8000/v1"
export OPENAI_API_KEY="EMPTY"

[!Tip]

We recommend using the following set of sampling parameters for generation

- Thinking mode for general tasks: temperature=1.0, top_p=0.95, top_k=20, min_p=0.0, presence_penalty=0.0, repetition_penalty=1.0

- Thinking mode for precise coding tasks (e.g. WebDev): temperature=0.6, top_p=0.95, top_k=20, min_p=0.0, presence_penalty=0.0, repetition_penalty=1.0

- Instruct (or non-thinking) mode: temperature=0.7, top_p=0.80, top_k=20, min_p=0.0, presence_penalty=1.5, repetition_penalty=1.0

Please note that the support for sampling parameters varies according to inference frameworks.

[!Important]

Qwen3.6 models operate in thinking mode by default, generating thinking content signified by \n...\n\n before producing the final responses.

To disable thinking content and obtain direct response, refer to the examples here.

Text-Only Input

from openai import OpenAI
# Configured by environment variables
client = OpenAI()

messages = [
    {"role": "user", "content": "Type \"I love Qwen3.6\" backwards"},
]

chat_response = client.chat.completions.create(
    model="Qwen/Qwen3.6-27B",
    messages=messages,
    max_tokens=81920,
    temperature=1.0,
    top_p=0.95,
    presence_penalty=0.0,
    extra_body={
        "top_k": 20,
    }, 
)
print("Chat response:", chat_response)

Image Input

from openai import OpenAI
# Configured by environment variables
client = OpenAI()

messages = [
    {
        "role": "user",
        "content": [
            {
                "type": "image_url",
                "image_url": {
                    "url": "https://qianwen-res.oss-accelerate.aliyuncs.com/Qwen3.5/demo/CI_Demo/mathv-1327.jpg"
                }
            },
            {
                "type": "text",
                "text": "The centres of the four illustrated circles are in the corners of the square. The two big circles touch each other and also the two little circles. With which factor do you have to multiply the radii of the little circles to obtain the radius of the big circles?\nChoices:\n(A) $\\frac{2}{9}$\n(B) $\\sqrt{5}$\n(C) $0.8 \\cdot \\pi$\n(D) 2.5\n(E) $1+\\sqrt{2}$"
            }
        ]
    }
]

response = client.chat.completions.create(
    model="Qwen/Qwen3.6-27B",
    messages=messages,
    max_tokens=81920,
    temperature=1.0,
    top_p=0.95,
    presence_penalty=0.0,
    extra_body={
        "top_k": 20,
    }, 
)
print("Chat response:", chat_response)

Video Input

from openai import OpenAI
# Configured by environment variables
client = OpenAI()

messages = [
    {
        "role": "user",
        "content": [
            {
                "type": "video_url",
                "video_url": {
                    "url": "https://qianwen-res.oss-accelerate.aliyuncs.com/Qwen3.5/demo/video/N1cdUjctpG8.mp4"
                }
            },
            {
                "type": "text",
                "text": "How many porcelain jars were discovered in the niches located in the primary chamber of the tomb?"
            }
        ]
    }
]

# When vLLM is launched with `--media-io-kwargs '{"video": {"num_frames": -1}}'`,
# video frame sampling can be configured via `extra_body` (e.g., by setting `fps`).
# This feature is currently supported only in vLLM.
#
# By default, `fps=2` and `do_sample_frames=True`.
# With `do_sample_frames=True`, you can customize the `fps` value to set your desired video sampling rate.
response = client.chat.completions.create(
    model="Qwen/Qwen3.6-27B",
    messages=messages,
    max_tokens=81920,
    temperature=1.0,
    top_p=0.95,
    presence_penalty=0.0,
    extra_body={
        "top_k": 20,
        "mm_processor_kwargs": {"fps": 2, "do_sample_frames": True},
    }, 
)

print("Chat response:", chat_response)

Instruct (or Non-Thinking) Mode

[!Important]

Qwen3.6 does not officially support the soft switch of Qwen3, i.e., /think and /nothink.

Qwen3.6 will think by default before response.

You can obtain direct response from the model without thinking by configuring the API parameters.

For example,

from openai import OpenAI
# Configured by environment variables
client = OpenAI()

messages = [
    {
        "role": "user",
        "content": [
            {
                "type": "image_url",
                "image_url": {
                    "url": "https://qianwen-res.oss-accelerate.aliyuncs.com/Qwen3.6/demo/RealWorld/RealWorld-04.png"
                }
            },
            {
                "type": "text",
                "text": "Where is this?"
            }
        ]
    }
]

chat_response = client.chat.completions.create(
    model="Qwen/Qwen3.6-27B",
    messages=messages,
    max_tokens=32768,
    temperature=0.7,
    top_p=0.8,
    presence_penalty=1.5,
    extra_body={
        "top_k": 20,
        "chat_template_kwargs": {"enable_thinking": False},
    }, 
)
print("Chat response:", chat_response)

[!Note]

If you are using APIs from Alibaba Cloud Model Studio, in addition to changing model, please use "enable_thinking": False instead of "chat_template_kwargs": {"enable_thinking": False}.

Preserve Thinking

By default, only the thinking blocks generated in handling the latest user message is retained, resulting in a pattern commonly as interleaved thinking.

Qwen3.6 has been additionally trained to preserve and leverage thinking traces from historical messages.

You can enable this behavior by setting the preserve_thinking option:

from openai import OpenAI
# Configured by environment variables
client = OpenAI()

messages = [...]

chat_response = client.chat.completions.create(
    model="Qwen/Qwen3.6-27B",
    messages=messages,
    max_tokens=32768,
    temperature=0.6,
    top_p=0.95,
    presence_penalty=0.0,
    extra_body={
        "top_k": 20,
        "chat_template_kwargs": {"preserve_thinking": True},
    }, 
)
print("Chat response:", chat_response)

[!Note]

If you are using APIs from Alibaba Cloud Model Studio, in addition to changing model, please use "preserve_thinking": True instead of "chat_template_kwargs": {"preserve_thinking": False}.

This capability is particularly beneficial for agent scenarios, where maintaining full reasoning context can enhance decision consistency and, in many cases, reduce overall token consumption by minimizing redundant reasoning. Additionally, it can improve KV cache utilization, optimizing inference efficiency in both thinking and non-thinking modes.

Agentic Usage

Qwen3.6 excels in tool calling capabilities.

Qwen-Agent

We recommend using Qwen-Agent to quickly build Agent applications with Qwen3.6.

To define the available tools, you can use the MCP configuration file, use the integrated tool of Qwen-Agent, or integrate other tools by yourself.

import os
from qwen_agent.agents import Assistant

# Define LLM
# Using Alibaba Cloud Model Studio
llm_cfg = {
    # Use the OpenAI-compatible model service provided by DashScope:
    'model': 'qwen3.6-27b',
    'model_type': 'qwenvl_oai',
    'model_server': 'https://dashscope.aliyuncs.com/compatible-mode/v1',
    'api_key': os.getenv('DASHSCOPE_API_KEY'),

    'generate_cfg': {
        'use_raw_api': True,
        # When using Dash Scope OAI API, pass the parameter of whether to enable thinking mode in this way
        'extra_body': {
            'enable_thinking': True,
            'preserve_thinking': True,
        },
    },
}

# Using OpenAI-compatible API endpoint.
# functionality of the deployment frameworks and let Qwen-Agent automate the related operations.
#
# llm_cfg = {
#     # Use your own model service compatible with OpenAI API by vLLM/SGLang:
#     'model': 'Qwen/Qwen3.6-27B',
#     'model_type': 'qwenvl_oai',
#     'model_server': 'http://localhost:8000/v1',  # api_base
#     'api_key': 'EMPTY',
#
#     'generate_cfg': {
#         'use_raw_api': True,
#         # When using vLLM/SGLang OAI API, pass the parameter of whether to enable thinking mode in this way
#         'extra_body': {
#             'chat_template_kwargs': {'enable_thinking': True, 'preserve_thinking': True}
#         },
#     },
# }

# Define Tools
tools = [
    {'mcpServers': {  # You can specify the MCP configuration file
            "filesystem": {
                "command": "npx",
                "args": ["-y", "@modelcontextprotocol/server-filesystem", "/Users/xxxx/Desktop"]
            }
        }
    }
]

# Define Agent
bot = Assistant(llm=llm_cfg, function_list=tools)

# Streaming generation
messages = [{'role': 'user', 'content': 'Help me organize my desktop.'}]
for responses in bot.run(messages=messages):
    pass
print(responses)

# Streaming generation
messages = [{'role': 'user', 'content': 'Develop a dog website and save it on the desktop'}]
for responses in bot.run(messages=messages):
    pass
print(responses)

Qwen Code

Qwen Code is an open-source AI agent for the terminal, optimized for Qwen models. It helps you understand large codebases, automate tedious work, and ship faster.

For more information, please refer to Qwen Code.

Processing Ultra-Long Texts

Qwen3.6 natively supports context lengths of up to 262,144 tokens.

For long-horizon tasks where the total length (including both input and output) exceeds this limit, we recommend using RoPE scaling techniques to handle long texts effectively., e.g., YaRN.

YaRN is currently supported by several inference frameworks, e.g., transformers, vllm, ktransformers and sglang.

In general, there are two approaches to enabling YaRN for supported frameworks:

Modifying the model configuration file:

In the config.json file, change the rope_parameters fields in text_config to:

    {
        "mrope_interleaved": true,
        "mrope_section": [
            11,
            11,
            10
        ],
        "rope_type": "yarn",
        "rope_theta": 10000000,
        "partial_rotary_factor": 0.25,
        "factor": 4.0,
        "original_max_position_embeddings": 262144,
    }

Passing command line arguments:

For vllm, you can use

    VLLM_ALLOW_LONG_MAX_MODEL_LEN=1 vllm serve ... --hf-overrides '{"text_config": {"rope_parameters": {"mrope_interleaved": true, "mrope_section": [11, 11, 10], "rope_type": "yarn", "rope_theta": 10000000, "partial_rotary_factor": 0.25, "factor": 4.0, "original_max_position_embeddings": 262144}}}' --max-model-len 1010000

For sglang and ktransformers, you can use

    SGLANG_ALLOW_OVERWRITE_LONGER_CONTEXT_LEN=1 python -m sglang.launch_server ... --json-model-override-args '{"text_config": {"rope_parameters": {"mrope_interleaved": true, "mrope_section": [11, 11, 10], "rope_type": "yarn", "rope_theta": 10000000, "partial_rotary_factor": 0.25, "factor": 4.0, "original_max_position_embeddings": 262144}}}' --context-length 1010000

[!NOTE]

All the notable open-source frameworks implement static YaRN, which means the scaling factor remains constant regardless of input length, potentially impacting performance on shorter texts.

We advise modifying the rope_parameters configuration only when processing long contexts is required.

It is also recommended to modify the factor as needed. For example, if the typical context length for your application is 524,288 tokens, it would be better to set factor as 2.0.

Best Practices

To achieve optimal performance, we recommend the following settings:

1. Sampling Parameters:

We suggest using the following sets of sampling parameters depending on the mode and task type:
Thinking mode for general tasks:

temperature=1.0, top_p=0.95, top_k=20, min_p=0.0, presence_penalty=0.0, repetition_penalty=1.0

Thinking mode for precise coding tasks (e.g., WebDev):

temperature=0.6, top_p=0.95, top_k=20, min_p=0.0, presence_penalty=0.0, repetition_penalty=1.0

Instruct (or non-thinking) mode:

temperature=0.7, top_p=0.80, top_k=20, min_p=0.0, presence_penalty=1.5, repetition_penalty=1.0

For supported frameworks, you can adjust the presence_penalty parameter between 0 and 2 to reduce endless repetitions. However, using a higher value may occasionally result in language mixing and a slight decrease in model performance.

2. Adequate Output Length: We recommend using an output length of 32,768 tokens for most queries. For benchmarking on highly complex problems, such as those found in math and programming competitions, we suggest setting the max output length to 81,920 tokens. This provides the model with sufficient space to generate detailed and comprehensive responses, thereby enhancing its overall performance.

3. Standardize Output Format: We recommend using prompts to standardize model outputs when benchmarking.

Math Problems: Include "Please reason step by step, and put your final answer within \boxed{}." in the prompt.
Multiple-Choice Questions: Add the following JSON structure to the prompt to standardize responses: "Please show your choice in the answer field with only the choice letter, e.g., "answer": "C"."

4. Long Video Understanding: To optimize inference efficiency for plain text and images, the size parameter in the released video_preprocessor_config.json is conservatively configured. It is recommended to set the longest_edge parameter in the video_preprocessor_config file to 469,762,048 (corresponding to 224k video tokens) to enable higher frame-rate sampling for hour-scale videos and thereby achieve superior performance. For example,

    {"longest_edge": 469762048, "shortest_edge": 4096}

Alternatively, override the default values via engine startup parameters. For implementation details, refer to: vLLM / SGLang.

Citation

If you find our work helpful, feel free to give us a cite.

@misc{qwen3.6-27b,
    title  = {{Qwen3.6-27B}: Flagship-Level Coding in a {27B} Dense Model},
    author = {{Qwen Team}},
    month  = {April},
    year   = {2026},
    url    = {https://qwen.ai/blog?id=qwen3.6-27b}
}

EXAMPLE GENERATION(s):

Some formatting may be lost on copy/paste.

Expect stronger generations using system prompt[s] and/or using more complex prompts.

Temp 1, rep pen 1 [off], topp .95 minp .05 topk 64

TWO EXAMPLE generations below, first one is 13k (!), second is "creative"

EXAMPLE #1

Sample Generation:

Q4_K_S (4-bit) ; non imatrix, standard default settings for Qwen 3.6 version model.

NOTES:

Some formatting lost on copy/paste ; original is flawless in terms of formatting, tables and charts.
Q4KS non-imatrix represents LOW end of the quality scale at 1/4 the precision of the full model.
Export of MTP generated content is broken in LMStudio (all formats)

PROMPT:

Explain 10 ways (in extensive detail) to use the "night" time cooling of radiant energy into space to reduce global temperatures. Include charts, and tables.

@page { size: 21cm 29.7cm; margin: 2cm }

p { line-height: 115%; margin-bottom: 0.25cm; background: transparent }

h1 { margin-bottom: 0.21cm; background: transparent; page-break-after: avoid }

h1.western { font-family: "Liberation Serif", serif; font-weight: bold; font-size: 24pt }

h1.cjk { font-size: 24pt; font-family: "NSimSun"; font-weight: bold }

h1.ctl { font-family: "Arial"; font-size: 24pt; font-weight: bold }

h2 { margin-top: 0.35cm; margin-bottom: 0.21cm; background: transparent; page-break-after: avoid }

h2.western { font-family: "Liberation Serif", serif; font-weight: bold; font-size: 18pt }

h2.cjk { font-size: 18pt; font-family: "NSimSun"; font-weight: bold }

h2.ctl { font-family: "Arial"; font-size: 18pt; font-weight: bold }

h3 { margin-top: 0.25cm; margin-bottom: 0.21cm; background: transparent; page-break-after: avoid }

h3.western { font-family: "Liberation Serif", serif; font-weight: bold; font-size: 14pt }

h3.cjk { font-size: 14pt; font-family: "NSimSun"; font-weight: bold }

h3.ctl { font-family: "Arial"; font-size: 14pt; font-weight: bold }

td p { orphans: 0; widows: 0; background: transparent }

th p { font-weight: bold; text-align: center; orphans: 0; widows: 0; background: transparent }

h4 { margin-top: 0.21cm; margin-bottom: 0.21cm; background: transparent; page-break-after: avoid }

h4.western { font-family: "Liberation Serif", serif; font-weight: bold; font-size: 12pt }

h4.cjk { font-size: 12pt; font-family: "NSimSun"; font-weight: bold }

h4.ctl { font-family: "Arial"; font-size: 12pt; font-weight: bold }

pre { background: transparent }

pre.western { font-family: "Liberation Mono", monospace; font-size: 10pt }

pre.cjk { font-size: 10pt; font-family: "NSimSun", monospace }

pre.ctl { font-family: "Liberation Mono", monospace; font-size: 10pt }

strong { font-weight: bold }

code.western { font-family: "Liberation Mono", monospace }

code.cjk { font-family: "NSimSun", monospace }

code.ctl { font-family: "Liberation Mono", monospace }

em { font-style: italic }

Thought

for 8.68 seconds

The

user wants me to explain 10 ways to use nighttime radiative cooling

(radiation of energy to space) to reduce global temperatures. This is

a fascinating topic combining atmospheric physics, climate

engineering, and engineering design.

Let

me think about what radiative cooling to space means:

night, especially under clear skies, surfaces can radiate thermal

energy directly to space through the "infrared window"

(roughly 8-13 μm wavelength range) where the atmosphere is

relatively transparent

This

causes surfaces to cool below ambient air temperature (radiative

cooling)

This

is why we see frost even when air temperatures are above freezing

Now

I need to think of 10 extensive approaches to leverage this

phenomenon for global temperature reduction:

Radiative

cooling materials for buildings and infrastructure -

Coatings, tiles, films that maximize emission in the IR window and

minimize solar absorption

Large-scale

radiative cooling panels/farms - Dedicated areas of

optimized radiative surfaces

Nighttime

ocean radiative cooling enhancement - Artificial surfaces

or materials in oceans to radiate heat at night

Atmospheric

water harvesting through radiative cooling - Condensing

moisture which has cooling effects and water benefits

Radiative

cooling towers - Structures similar to cooling towers but

using radiative cooling to chill water for district cooling or

direct injection

Selective

surface radiators in the upper atmosphere - Balloons or

structures at high altitude where radiative cooling is more

effective

Enhanced

desert radiative cooling - Modifying desert surfaces to

maximize nighttime heat loss

Radiative

cooling combined with heat pumps - Using radiative cooling

as heat sink for heat pumps at night

Nighttime

urban radiative cooling corridors - Designing urban areas

to channel and enhance radiative cooling effects

Radiative

cooling for carbon capture enhancement - Using radiative

cooling surfaces to condense CO2 or enhance carbon capture systems

I'll

develop these concepts with detailed scientific analysis, supporting

visual representations, and practical implementation strategies for

each approach.

OUTPUT:

Radiative

Nighttime Cooling for Global Temperature Reduction: Ten Comprehensive

Approaches

Introduction

to Radiative Cooling to Space

Radiative

cooling occurs when a surface emits thermal radiation to the cold

sky, particularly through the atmospheric "infrared window"

(8–13 μm) where greenhouse gases are relatively transparent. At

night, under clear conditions, surfaces can lose more energy than

they gain, cooling below ambient air temperature—sometimes by

10–15°C or more. This natural phenomenon, responsible for

nighttime frost, can be engineered into systematic climate

intervention strategies.

This

document examines ten approaches to harness this effect at scale for

global temperature mitigation.

Radiative Cooling Building Envelopes and Infrastructure

Principle

and Physics

Buildings

and paved surfaces constitute ~40% of the urban surface area

globally. Replacing conventional materials with selective radiative

cooling coatings that maximize emission in the 8–13 μm atmospheric

window while minimizing solar absorption can transform urban areas

from heat sources to heat sinks.

The

net radiative heat flux from a surface is:

Q_net

= ε_surface · σ · T_surface⁴ - Q_solar(1-α) - Q_downward_IR -

Q_convective - Q_conductive

Where:

ε_surface

= emissivity of surface in IR window

= Stefan-Boltzmann constant (5.67×10⁻⁸ W/m²/K⁴)

= solar reflectivity

Q_downward_IR

= atmospheric downwelling IR radiation

Q_convective/conductive

= sensible heat transfer to/from surroundings

Materials

and Engineering

Optimal

coating properties:

Emissivity

>0.90 in 8–13 μm band

Solar

reflectivity >0.85

Durability:

>20 years outdoor exposure

Cost:

<$5/m² at scale

Current

material classes:

Material Class

IR Emissivity (8-13μm)

Solar Reflectivity

Nighttime ΔT (°C)

Cost ($/m²)

High-performance polymer

coatings

0.92

0.88

-14

8-15

Spectrally selective

paints (SiO₂/MgO)

0.88

0.82

-10

3-6

Metamaterial photonic

coolers

0.95

0.91

-17

50+

Bare concrete (control)

0.85

0.40

-4

Bitumen asphalt

(control)

0.90

0.08

Implementation

Scale

Global

urban area breakdown:

Roofs:

~100 million km²

Walls:

~150 million km²

Pavements:

~180 million km²

Total

addressable: ~430 million km²

Phase

1 (2025-2035): 10% coverage (43 million km²)

Focus:

New construction mandates, retrofit high-value infrastructure

Investment:

$200-400 billion

Phase

2 (2035-2050): 30% coverage (129 million km²)

Include:

Residential retrofitting programs, public infrastructure

Cumulative

investment: $1-1.5 trillion

Phase

3 (2050-2075): 50-70% coverage

Global

standard building codes

Cumulative

investment: $2-3 trillion

Cooling

Performance

Nighttime

radiative cooling power by latitude:

Latitude Band

Clear Sky Nights/Year

Avg Radiative Power (W/m²)

Max ΔT vs Ambient (°C)

0-20° (Tropical)

180-250

45-65

-8 to -12

20-40° (Subtropical)

200-280

55-80

-12 to -16

40-60° (Temperate)

150-220

60-90

-10 to -14

60-90° (Boreal)

100-180

40-70

-6 to -10

Global

Impact Estimate

Assuming

40% global urban coverage with advanced radiative cooling materials:

Average

nighttime cooling power: 50 W/m² effective

Area:

172 million km² = 1.72×10¹⁴ m²

Night

hours/year: ~4,380 hours = 1.58×10⁷ s

Total

annual energy diverted: 50 × 1.72×10¹⁴ × 1.58×10⁷ =

1.36×10²³ J = 136 exajoules/year

Equivalent

to removing ~500 million metric tons CO₂ equivalent per year

Challenges

Humidity/fog

reduces IR transmission through atmospheric window

Wind

increases convective heat transfer, offsetting radiative cooling

Long-term

material durability and maintenance

Requires

policy mandates for new construction and retrofits

Dedicated Radiative Cooling Farms

Concept

Purpose-built

facilities consisting of large, optimized radiative cooling surfaces

designed explicitly for climate cooling rather than as building

byproducts. These "cooling farms" would be situated in arid

or semi-arid regions where nighttime clear skies are prevalent.

System

Design

Component

layout:

Radiative

panels: 80% of surface area

Support

structures: 10%

Monitoring

and control: 5%

Access

and infrastructure: 5%

Panel

specifications:

Parameter

Specification

Panel dimensions

10 m × 5 m (50 m²

each)

Tilt angle

0-45° (optimizable by

latitude)

Surface material

SiO₂/MgO composite

coating on aluminum substrate

Spacing

2 m minimum (air

circulation)

Thermal mass

Minimized (rapid nightly

cooling)

Heat

Transfer Mechanisms

Three

primary pathways for heat removal:

Direct

radiative loss to space (primary, 60-70%)

Convective

transfer to night air, then vertical mixing (20-30%)

Conductive

transfer to ground (variable, minimized)

Radiative

heat loss calculation (typical clear night):

Surface

temperature: 278 K (5°C) Surface emissivity in window: 0.92

Effective sky temperature (clear desert night): 258 K (-15°C)

Q_rad

= ε · σ · (T_surface⁴ - T_sky⁴) Q_rad = 0.92 ×

5.67×10⁻⁸ × (278⁴ - 258⁴) Q_rad ≈ 78 W/m²

Site

Selection Criteria

Optimal

characteristics:

Low

humidity (<40% RH at night)

High

nighttime cloud cover probability (>70%)

Low

wind speed at night (<5 m/s average)

Flat

terrain

Non-agricultural

land

Access

to monitoring infrastructure

Global

candidate regions:

Region

Area Available (km²)

Annual Clear Night Hours

Avg Cooling Power (W/m²)

Sahara Desert

500,000

2,800

Arabian Desert

150,000

2,600

Australian Outback

400,000

2,400

Gobi Desert

80,000

2,200

Great Basin (USA)

30,000

2,000

Patagonia

50,000

1,800

Total

1,240,000

~2,200

~55

Scale

and Economics

Single

facility (100 km²):

Metric

Value

Radiative panel area

80 km² = 8×10⁷ m²

Avg nighttime cooling

power

50 W/m²

Annual heat diverted

8.8×10¹⁶ J/year (88

PJ)

Initial capital cost

$400 million

O&M annual cost

$4 million

Cost per ton CO₂eq

$5-10/ton-year

Global

deployment scenario:

Phase

Facilities

Total Area (km²)

Annual Heat Diverted (EJ)

Cumulative Cost (B$)

1 (pilot)

1,000

8.8

2 (scale)

100

10,000

3 (regional)

500

50,000

440

250

4 (global)

2,000

200,000

1,760

1,000

Climate

Impact

Phase 4 deployment (200,000 km²):

Annual

heat diverted: 1,760 EJ

CO₂

equivalent: ~7,000 metric tons/year

Estimated

global temperature effect: 0.01-0.03°C

While

modest alone, radiative cooling farms provide:

Zero

operational emissions

Potential

co-production of water (condensation)

Synergistic

use with solar PV (daytime solar, nighttime cooling)

Demonstrable,

measurable effects for monitoring

Nighttime Ocean Radiative Cooling Enhancement

Background

The

oceans cover ~71% of Earth's surface and store ~90% of excess heat

from greenhouse warming. Nighttime radiative cooling of the ocean

surface naturally occurs but is limited by:

High

evaporative loss (latent heat transfer upward)

Turbulent

mixing bringing warmer water from below

Cloud

cover reducing clear-sky conditions

Enhancement

Strategies

Three

complementary approaches:

Surface Microlayer Enhancement

Deploy

biodegradable, IR-transparent, solar-reflective materials that form a

thin layer on the ocean surface:

Material

requirements:

Low

thermal conductivity (reduce mixing with subsurface water)

High

IR emissivity in 8-13 μm window

High

solar reflectivity

Biodegradable

within 24-72 hours

Example:

Polymer microsphere layer

Composition:

Silica or polyurethane microspheres

Layer

thickness: 10-50 μm

Buoyant

and self-arranging

Washed

off naturally by waves

Artificial Ice/Brine Formation

polar and subpolar regions, induce formation of thin ice or

concentrated brine layers at night that:

Have

lower thermal conductivity than water

Radiate

more efficiently to space

Melt

during daytime (no permanent accumulation)

Subsurface Upwelling at Night

Use

pumps or mixing devices to bring colder subsurface water to the

surface at night when radiative cooling is most effective, then allow

mixing back during daytime.

Energy

Balance Analysis

Current

ocean nighttime heat budget (per m²):

Heat Losses:

├── Radiative loss to space (clear sky) 40-60 W/m²

├── Evaporative loss 20-50 W/m²

└── Convective loss 5-15 W/m²

Heat Gains:

├── Downwelling IR from atmosphere 30-50 W/m²

├── Heat from subsurface mixing 10-30 W/m²

└── Upwelling from depth variable

Net:

Often slightly positive (ocean gains heat) due to mixing and

evaporation

With

enhancement (microlayer approach):

Modified budget:

├── Radiative loss to space (enhanced) 60-80 W/m² (+20-30%)

├── Evaporative loss (reduced) 5-15 W/m² (-60-70%)

├── Convective loss 5-10 W/m²

├── Downwelling IR (unchanged) 30-50 W/m²

└── Heat from mixing (reduced) 2-8 W/m² (-70-80%)

Net:

15-35 W/m² heat LOSS to atmosphere/space

Implementation

Infrastructure

Microlayer

deployment system:

Component

Description

Carrier vessels

Modified tankers,

autonomous surface vehicles

Distribution

Boom spreaders, spray

systems

Target areas

5-15 km² patches

Frequency

Daily (material

biodegrades)

Cost

$50-200 per km² per day

Seasonal

deployment strategy:

Region

Active Months

Rationale

Arctic

May-September

Maximum daylight/heat

gain period

Subarctic (N)

June-September

Peak warming

Subarctic (S)

December-March

Peak warming

Tropical Pacific

Year-round

Consistent conditions

Global

Potential

Assuming

1% of ocean surface treated (3.6 million km²):

Average

nighttime cooling power: 20 W/m² net

Night

hours/year: 4,380 hours

Annual

heat removed: 20 × 3.6×10¹² × 1.58×10⁷ = 1.14×10²¹ J =

1,140 EJ/year

CO₂

equivalent removal: ~4,500 metric tons/year

Risks

and Considerations

Ecosystem

impact: Potential effects on marine organisms,

especially plankton and larval stages

Material

accumulation: Risk of microplastic pollution if

biodegradation fails

Altered

evaporation: Changes to precipitation patterns

Economic

viability: High ongoing costs for material production

and deployment

Regulatory

complexity: International waters governance

Atmospheric Water Harvesting via Radiative Cooling

Mechanism

Radiative

cooling surfaces that drop below the dew point of ambient air

condense water vapor into liquid water. This process is both a water

resource and a cooling mechanism:

Radiative

surface cooling - Surface cools below ambient via IR

emission

Condensation -

Water vapor condenses on cold surface

Latent

heat release - Released heat is radiated away during

continued nighttime cooling

Cooling

effect - Surface stays cooler than it otherwise would

due to evaporative/latent cooling cycle

System

Design

Atmospheric

water harvester (AWH) with climate cooling function:

Component

Specification

Condensing surface

Copper or aluminum with

hydrophobic coating

Surface area per unit

50-500 m²

Target temperature

10-15°C below ambient

Collection system

Tilted panels to

channels

Storage

Insulated tanks

Power requirement

Minimal (fans, pumps

optional)

Performance

by humidity:

RH (%)

Temp (°C)

Dew Point (°C)

Water Yield (L/m²/night)

Latent Heat Released (kJ/m²)

-6

1.2

2,700

3.5

8,000

5.0

11,500

Dual-Benefit

Analysis

For

each liter of water harvested:

Water

produced: 1 L (value: $0.10-$10 depending on location)

Cooling

effect:

Latent

heat of vaporization: 2,260 J/g

Per

liter: 2.26 MJ of heat moved from air to surface and radiated away

Extended

radiative cooling: Wet surfaces can radiate more

effectively than dry ones

Climate

vs. water benefits by region:

Region

Annual Water Yield (L/m²)

Annual Cooling (MJ/m²)

Primary Benefit

Coastal California

2,500

180

Water

Middle East

1,800

130

Water + cooling

Northern Africa

2,200

160

Water

South Asia

4,000

290

Cooling + water

Southeast Asia

5,000

365

Cooling

Large-Scale

Deployment

Urban

integration approach:

Rooftop

AWH systems in coastal and semi-arid cities

Integration

with building cooling systems

Scale:

1 m² AWH per 10 m² of building

Global

potential (assuming 50 million m² total AWH area in arid/semi-arid

regions):

Annual

water production: ~250 million L

Annual

heat radiated away via latent heat mechanism: ~1.8×10¹⁴ J

Additional

heat from enhanced radiative surface cooling: ~3.6×10¹⁴ J

Total

cooling: ~5.4×10¹⁴ J/year (0.54 EJ)

CO₂

equivalent: ~2,000 metric tons/year

Advantages

Addresses

two climate challenges simultaneously (water scarcity + warming)

Passive

operation with minimal energy

Synergistic

with building energy efficiency

Water

can support vegetation, further cooling via transpiration

Radiative Cooling Towers

Concept

Massive

structures analogous to industrial cooling towers, but designed to

radiate heat directly to space rather than using evaporative cooling.

These towers maximize surface area-to-volume ratio for radiative loss

and are positioned to access cooler nighttime air.

Engineering

Design

Tower

geometry:

Hyperboloid

shape (similar to existing cooling towers)

Height:

100-300 m

Base

diameter: 150-400 m

Top

diameter: 50-150 m

Surface

treatment:

Entire

interior and exterior coated with high-emissivity,

high-solar-reflectivity materials

Surface

area per tower: 50,000-300,000 m²

Heat

transfer modes within tower:

Air-borne

heat removal (natural convection):

Warm

air rises through tower, cooling via contact with radiating walls

Heat

radiated from walls to night sky

Cooled

air exits at top and disperses

Liquid-borne

heat removal (optional):

Warm

water circulated through tower exterior/interior

Water

cooled radiatively, then pumped back to source

Can

serve district cooling applications

Performance

Calculations

Radiative

cooling tower (200 m tall, 200 m base diameter):

Parameter

Value

Surface area

180,000 m²

Effective emissivity

0.88

Average night

temperature

15°C

Effective sky

temperature (clear)

-10°C

Radiative power per m²

~55 W/m²

Total radiative cooling

power

9.9 MW

Annual heat removed

(clear nights)

1.2×10¹¹ kJ

Comparison

to conventional evaporative tower:

Metric

Radiative Tower

Evaporative Tower

Cooling capacity

10-20 MW

50-100 MW

Water usage

0 L/h

5,000-10,000 L/h

Energy input

0-50 kW

500 kW-2 MW

Nighttime efficiency

100% (passive)

70-90%

Climate benefit

Direct + no emissions

Direct only

Heat

Sink Applications

Three

primary use cases:

Nighttime Urban Heat Disposal

Collect

heat from urban buildings during day (via district heating/thermal

storage)

Radiate

it away at night through cooling towers

Reduces

daytime air conditioning demand

Power Plant Heat Sink

Replace

or supplement evaporative cooling at thermal/nuclear plants

Particularly

valuable in water-scarce regions

Direct Climate Cooling

Towers

designed solely to radiate ambient heat to space

Positioned

in high-altitude, clear-sky regions

Global

Deployment Scenario

Phase

1: 100 radiative cooling towers

Locations:

Major urban centers (2-5 per city)

Total

annual heat removed: 1.2×10¹³ kJ

Phase

2: 1,000 towers

Expanded

urban and industrial coverage

Total

annual heat removed: 1.2×10¹⁴ kJ

Phase

3: 5,000 towers

Global

coverage of major population/industrial centers

Total

annual heat removed: 6×10¹⁴ kJ (0.6 EJ)

Cost

estimates:

Construction

per tower: $50-200 million

Phase

3 total construction: $250-1,000 billion

CO₂

equivalent removal (Phase 3): ~2,500 metric tons/year

Technical

Challenges

Lower

cooling capacity than evaporative towers (must overcome with scale)

High

construction cost per unit

Requires

clear-sky regions for optimal performance

Wind

loads and structural design at large heights

Upper-Altitude Radiative Cooling Platforms

Principle

high altitudes, the atmospheric density is lower, providing less

obstruction to radiative cooling. Platforms (balloons, gliders, or

satellites) carrying radiative cooling surfaces at 20-50 km altitude

can radiate heat directly to space with minimal atmospheric

interference.

Platform

Types

High-Altitude Balloons

Characteristics:

Operating

altitude: 20-35 km

Duration:

Weeks to months

Radiative

surface area per balloon: 50-500 m²

Power:

Solar PV for station-keeping and telemetry

Advantages:

Low

cost compared to satellites

Easy

to deploy and replace

Access

to mesosphere where IR window is nearly fully open

Aerostats (Buoyant Platforms)

Characteristics:

Operating

altitude: 15-30 km

Duration:

Years

Radiative

surface area per platform: 500-5,000 m²

Power:

Solar + batteries

Low-Earth Orbit Satellites

Characteristics:

Altitude:

200-800 km

Radiative

surface area per satellite: 1,000-10,000 m²

atmospheric obstruction

Continuous

radiative cooling (except during eclipse)

Radiative

Cooling Performance vs. Altitude

Effective

radiative cooling power:

Altitude

Atmospheric Pressure

Clear-Sky Factor

Effective T_sky (K)

Radiative Power (W/m²)

0 km (surface)

1013 mbar

0.70

248

10 km

265 mbar

0.92

215

20 km

55 mbar

0.98

185

35 km

12 mbar

1.00

165

110

50+ km (space)

~0 mbar

1.00

4.2 K

350+

System

Design: High-Altitude Balloon Fleet

Single

balloon system:

Parameter

Specification

Balloon type

Superpressure helium

Operating altitude

30 km

Radiative surface

200 m² of photonic

metamaterial

Radiative power

110 W/m² × 200 = 22 kW

Lifetime

90 days

Cost (including

deployment)

$500,000

Fleet

of 10,000 balloons (rotated continuously):

Active

at any time: ~5,000

Total

radiative power: 5,000 × 22 kW = 110 MW

Annual

heat radiated: 110 MW × 3.15×10⁷ s = 3.47×10¹⁵ J = 3,470 GJ

Annual

operational cost (replacements, telemetry): $2.5 billion

Satellite-Based

System

Constellation

design:

Parameter

Specification

Number of satellites

500

Orbit

600 km, sun-synchronous

Radiative surface per

sat

2,000 m² deployable

Radiative power per sat

350 W/m² × 2,000 = 700

Constellation power

350 MW

Lifetime

15 years

Cost per satellite

$100 million

Total constellation cost

$50 billion

Annual

heat radiated: 350 MW × 3.15×10⁷ s = 1.10×10¹⁶ J =

11 PJ/year

Cost-Effectiveness

Comparison

Platform

Cost per TJ Radiated

Operational Lifetime

Notes

Surface radiative cooler

$0.10

20+ years

Lowest cost

Cooling tower

$0.50

50+ years

Large-scale

High-altitude balloon

$8.00

90 days

Moderate

Aerostat

$5.00

5-10 years

Medium

LEO satellite

$4.50

15 years

Highest power

Strategic

Value

High-power

radiative cooling: Space-based platforms can radiate

5-10× more per m² than surface systems

Geographic

flexibility: Can target specific latitudes/longitudes

land use: Eliminates terrestrial ecological concerns

Dual-use

potential: Platforms could also monitor climate or

provide communications

Scalable: Start

small, expand incrementally

Challenges

High

cost per unit

Space

debris and orbital congestion concerns

Complex

launch and maintenance infrastructure

Political/regulatory

complexity for space-based climate engineering

Single-point

failure risk for satellites

Enhanced Desert Radiative Cooling

Background

Desert

regions naturally experience extreme nighttime radiative cooling due

to clear skies, low humidity, and minimal vegetation. However,

natural desert surfaces (sand, rock) have suboptimal radiative

properties and can be engineered to enhance this natural phenomenon.

Enhancement

Approaches

Surface Modification

Materials

and treatments:

Spread

high-emissivity mineral coatings (e.g., MgO, SiO₂) over desert

floors

Install

radiative cooling panels interspersed with natural terrain

Create

reflective gravel or stone pavements with high IR emissivity

Desert Radiative Corridors

Long,

narrow channels or "corridors" oriented to maximize IR

transmission to space:

Width:

50-200 m

Length:

10-100 km

Treated

surfaces on sides and floor

Oriented

perpendicular to prevailing night winds

Thermal Mass Reduction

Reduce

thermal mass of desert surfaces to enable deeper nighttime cooling:

Remove

or replace high-thermal-mass rocks and concrete

Install

lightweight radiative materials

Create

air gaps beneath surface layers

Quantitative

Analysis

Natural

desert night cooling vs. enhanced:

Parameter

Natural Desert Sand

Enhanced (SiO₂ coating)

Surface emissivity

(8-13μm)

0.82

0.93

Solar reflectivity

0.25

0.65

Thermal mass (J/kg·K)

800

350

Nighttime ΔT vs air

(°C)

-6 to -10

-14 to -20

Radiative power (W/m²)

35-50

60-85

Example:

Enhanced radiative cooling in Sahara

Area

treated: 100,000 km² (10% of Sahara)

Enhancement:

+30 W/m² average nighttime cooling power

Night

hours/year: 2,800 hours

Annual

additional heat radiated: 30 × 10¹¹ × 10⁴ × 10,080 = 3.0×10²⁰

J = 300 EJ

Infrastructure

and Economics

Treatment

methods:

Method

Cost ($/km²)

Lifetime

Maintenance

Mineral coating spread

500,000

5-10 years

Annual reapplication

Panel installation

2,000,000

20+ years

Low

Gravel paving

800,000

30+ years

Low

Air-gap substrate

1,200,000

25+ years

Medium

ROI

considerations:

direct economic return (pure climate benefit)

Potential

co-benefits: reduced daytime heating (less energy for cooling),

increased fog/condensation capture

Carbon

credit revenue possible under future markets

Regional

Climate Effects

Potential

secondary effects of large-scale desert radiative cooling:

Altered

wind patterns: Enhanced cooling could strengthen

nighttime thermal winds

Precipitation

changes: Cooler air holds less moisture, potentially

reducing fog/precipitation locally

Dust

reduction: Treated surfaces may reduce dust generation

Biodiversity

impacts: Temperature changes could affect desert

flora/fauna

Albedo

change: Increased reflectivity during daytime could

further reduce warming

Radiative Cooling as Heat Sink for Heat Pumps

Principle

Radiative

cooling surfaces can serve as the "cold side" (heat sink)

for heat pumps, enabling heat to be pumped from warm sources

(buildings, industrial processes, or the atmosphere) to the cold

night sky. This amplifies the natural radiative cooling effect

through active thermodynamic work.

System

Architecture

Nighttime

heat pump cycle:

[Heat Pump System]

┌──────────────┐

Warm Source (T_warm)│ │ Radiative Cooler (T_cold)

(e.g., │ Heat │ ──────► Night Sky (T_sky)

building, │ Pump │ (radiative loss)

process, │ (COP) │

air) │ │

└──────────────┘

▲

│

Electrical Power

Key

performance metric: Coefficient of Performance (COP)

COP

= Q_cooling / W_electrical

Where:

Q_cooling

= heat removed from warm source

W_electrical

= electrical power consumed

Technical

Specifications

Heat

pump coupled with radiative cooler:

Parameter

Value

Warm source temperature

20-30°C

Radiative cooler

temperature (night)

-5 to 10°C

Temperature lift (ΔT)

25-35 K

Heat pump type

Scroll or screw

compressor

COP (at design point)

2.5-4.0

Heat pump capacity

100 kW - 5 MW

Annual operating hours

1,500-2,500

Cooling

Power Amplification

Example:

1 MW heat pump coupled to 50,000 m² radiative cooler:

Parameter

Calculation

Result

Radiative cooler area

50,000 m²

Radiative power density

60 W/m²

Passive radiative

cooling

50,000 × 60

3 MW

Heat pump capacity

1 MW (electrical input)

COP

3.0

Active heat pumping

1 × 3.0

3 MW

Total

heat radiated

3 + 3

Result:

2× amplification over passive radiative cooling alone

Applications

District Cooling Systems

Collect

heat from buildings during daytime

Store

thermally (in water tanks, phase-change materials)

Radiate

away at night via heat pump + radiative cooler system

Reduces

or eliminates need for vapor-compression chillers

Annual

savings estimate (single urban district):

Replaces

5,000 tons of vapor-compression cooling

Eliminates

20 GWh/year electrical consumption

Reduces

CO₂ emissions: ~10,000 metric tons/year

Industrial Process Cooling

Industries

with nighttime-dominant cooling needs

Chemical

processing, food processing, pharmaceuticals

Potential

30-50% reduction in cooling costs vs. conventional chillers

Direct Climate Cooling

Heat

pumps extract heat from ambient air

Radiate

away at night via large-scale radiative coolers

Net

cooling of local and potentially regional atmosphere

Economics

Cost

analysis for 1 MW heat pump + radiative cooler system:

Component

Cost ($)

Radiative cooler (50,000

m²)

250,000

Heat pump (1 MW)

500,000

Electrical systems

100,000

Installation

150,000

Total

$1,000,000

Operating

costs (annual):

Electricity:

1,000 hours × 1 MW × $0.10/kWh = $100,000

O&M:

5% of capex = $50,000

Total

annual cost: $150,000

Cost

per ton CO₂ avoided (assuming 2,000 tons/year): $75/ton

Global

Potential

Assuming

10,000 systems deployed globally:

Total

heat radiated annually: 5×10²⁰ J = 500 EJ

CO₂

equivalent reduction: ~2,000 metric tons/year

Total

investment: $10 trillion

Advantages

Dramatically

amplifies natural radiative cooling effect

Provides

economic benefits (reduced cooling costs)

Can

be integrated into existing infrastructure

Scalable

from individual buildings to industrial complexes

Limitations

Requires

electrical power input

COP

decreases with larger temperature lifts

Highest

efficiency only during nighttime clear-sky conditions

Capital-intensive

initial deployment

Nighttime Urban Radiative Cooling Corridors

Concept

Design

urban environments to channel, preserve, and amplify nighttime

radiative cooling through engineered "cooling corridors"

that connect areas of high radiative cooling (parks, water bodies,

radiative cooling installations) throughout urban centers.

Urban

Heat Island Context

Urban

areas are typically 2-10°C warmer than surrounding rural areas due

to:

High

thermal mass of buildings and pavement

Waste

heat from vehicles and buildings

Reduced

vegetation and green space

Geometric

"canyon" effects trapping heat

Nighttime

heat budget of urban area (per m²):

Heat Sources:

├── Building waste heat 20-50 W/m²

├── Vehicle exhaust 5-15 W/m²

├── Ground heat release (thermal mass) 10-30 W/m²

└── Downwelling IR 30-50 W/m²

Heat Losses:

├── Radiative loss to sky 20-40 W/m² (limited by geometry)

├── Convective loss to air 10-20 W/m²

└── Lateral diffusion 5-15 W/m²

Net:

Typically positive (heat accumulates)

Corridor

Design Principles

Key

design features:

Geometric

alignment: Orient corridors perpendicular to

prevailing nighttime wind directions to channel cool air

Surface

treatment: Replace high-thermal-mass surfaces with

radiative cooling materials

Height-to-width

ratio: Maintain H/W < 0.5 to maximize sky view

factor and radiative loss

Barrier

removal: Eliminate obstacles that block airflow and

radiative heat loss

Connectivity: Link

corridors to form network rather than isolated features

Corridor

types:

Type

Width

Length

Sky View Factor

Primary Function

Street corridor

20-50 m

1-10 km

0.3-0.6

Airflow + radiation

Green corridor

50-200 m

5-20 km

0.7-0.9

Radiation +

evapotranspiration

Water corridor

10-100 m

1-50 km

0.9-1.0

Radiation + water

cooling

Rail corridor

30-100 m

10-100 km

0.6-0.8

Long-distance transport

Implementation

Strategy

Phase

1: Identify and map

Map

existing cooling sources (parks, water, radiative cooling

facilities)

Identify

wind corridors and airflow pathways

Locate

heat hotspots and areas needing cooling

Phase

2: Corridor creation

Retrofit

streets, parks, and waterways with radiative cooling surfaces

Remove

or modify barriers to airflow

Install

radiative cooling infrastructure along corridors

Phase

3: Network integration

Connect

corridors into cohesive urban-scale system

Implement

building facade treatments along corridor edges

Coordinate

with urban planning and zoning

Performance

Modeling

Cooling

effect of urban radiative corridor (100 m wide, 5 km long):

Parameter

Before Enhancement

After Enhancement

Surface emissivity

0.65

0.90

Sky view factor

0.40

0.75

Radiative power (W/m²)

Surface ΔT vs ambient

(°C)

-2

-8

Air temperature

reduction along corridor (°C)

-1 to -3

Annual

cooling energy (per corridor):

Enhanced

radiative power: 30 W/m² × 5×10⁵ m² = 15 MW

Night

hours/year: 2,000 hours

Annual

heat removed: 15 MW × 2,000 h = 30 GWh = 1.08×10¹¹ kJ

Urban-Scale

Deployment

For

a city of 1 million people (~100 km² urban area):

Infrastructure

Quantity

Area (km²)

Annual Heat Removed

Street corridors

1.0

5.4×10¹¹ kJ

Green corridors

0.5

2.7×10¹¹ kJ

Water/rail corridors

0.25

1.35×10¹¹ kJ

Total

1.75

9.45×10¹¹

Co-Benefits

Reduced

air conditioning demand: Cooler nighttime temperatures

reduce next-day cooling needs

Improved

air quality: Better ventilation reduces pollutant

concentrations

Biodiversity: Green

corridors provide habitat corridors

Flood

management: Enhanced drainage through green corridors

Public

health: Reduced heat-related mortality

Urban

resilience: Better adaptation to climate change

Economic

Analysis

Investment

per city (1 million population):

Component

Cost ($)

Street corridor

retrofitting

250,000,000

Green corridor creation

150,000,000

Water/rail corridor

enhancement

75,000,000

Radiative cooling

installations

100,000,000

Monitoring and control

25,000,000

Total

$600,000,000

Annual

benefits:

Energy

savings (reduced cooling): $50-100 million

Health

cost reduction: $10-30 million

Air

quality improvements: $5-15 million

Property

value increases: $20-50 million

Total

annual benefit: $85-195 million

Payback

period: 3-7 years (including climate benefits)

10.

Radiative Cooling for Carbon Capture Enhancement

Principle

Radiative

cooling can enhance carbon dioxide capture from atmospheric or flue

gas streams through:

Direct

condensation: Cooling surfaces below CO₂ dew point

to precipitate solid CO₂

Indirect

enhancement: Cooling gases to increase efficiency of

absorption/sorption processes

Hybrid

systems: Combining radiative cooling with other

capture technologies

Technical

Approaches

Direct CO₂ Condensation

Physical

requirements:

Temperature

below CO₂ sublimation point: -78.5°C (at 1 atm)

This

requires temperatures far below typical radiative cooling can

achieve alone

Practical

approach: Use radiative cooling as pre-cooling stage, then

mechanical refrigeration to reach sublimation point.

Radiative Cooling-Enhanced Amine Absorption

Amine

solvents absorb CO₂ more efficiently at lower temperatures.

Radiative cooling can:

Pre-cool

incoming gas stream

Cool

the amine solvent during absorption

Reduce

energy needed for solvent regeneration

Process

flow:

Flue Gas (hot, CO₂-rich)

│

▼

┌─────────────┐

│ Radiative │

│ Cooler │ ───────► Heat radiated to night sky

│ Pre-cooler │

└─────────────┘

│

▼

Cooler Flue Gas

│

▼

┌─────────────┐

│ Amine │

│ Absorber │

└─────────────┘

│

Clean Gas Exit

│

CO₂-rich Amine (for regeneration)

Membrane Separation Enhancement

Gas

separation membranes often perform better at lower temperatures for

CO₂. Radiative cooling can:

Reduce

membrane operating temperature

Increase

selectivity and permeation for CO₂

Reduce

compressor energy

Performance

Analysis

Radiative

cooling pre-cooling stage for flue gas capture:

Parameter

Without Pre-Cooling

With Radiative Pre-Cooling

Flue gas inlet temp

120°C

60°C

Amine absorption

efficiency

85%

93%

Regeneration energy

(GJ/ton CO₂)

3.5

3.0

CO₂ captured per kg

amine (kg/kg)

0.55

0.65

Energy savings

14%

Quantitative

example (power plant, 500 MW):

Flue

gas flow: 1.5×10⁶ kg/h

CO₂

concentration: 15%

Annual

CO₂ production: 1.99 million tons

Radiative

cooling contribution: 14% energy reduction in capture

Annual

energy savings: 3,486 GJ

CO₂

savings (via reduced capture energy): ~800 tons/year

System

Design

Radiative

cooling heat exchanger for gas cooling:

Parameter

Specification

Type

Plate-and-frame or

fin-tube heat exchanger

Radiative surface area

1,000-10,000 m²

Gas flow rate

1,000-50,000 kg/h

Temperature drop

20-40°C

Surface material

Aluminum with radiative

coating

Operating hours

Nighttime only (12

h/day)

Pressure drop

<50 mbar

Economic

Analysis

Radiative

cooling addition to carbon capture system (per 500 MW power plant):

Item

Cost ($)

Radiative cooling heat

exchanger

5,000,000

Installation

1,500,000

Controls and

instrumentation

500,000

Total

capital cost

$7,000,000

Annual

savings:

Energy

savings: $350,000/year

CO₂

credits (at $50/ton, 800 tons): $40,000/year

Total

annual benefit: $390,000

Simple

payback period: 18 years

Scaling

and Impact

For

10,000 power plants globally (assuming 10% adoption):

1,000

plants with radiative cooling enhancement

Annual

CO₂ capture energy savings: 800,000 tons CO₂ equivalent

Total

investment: $7 billion

Cost

per ton CO₂ saved: $8,750/ton (initial year)

Over

20-year lifetime: $438/ton

Limitations

Provides

modest improvements to carbon capture efficiency

Nighttime-only

operation limits applicability

Capital

costs may not be justified for small facilities

Climate

benefit is indirect (via improved capture efficiency) rather than

direct

Strategic

Value

Complementary

technology that enhances existing carbon capture

additional operational emissions

Can

be retrofitted to existing capture systems

Demonstrable

energy cost savings

Synergistic

with other nighttime radiative cooling applications

Comparative

Analysis: All Ten Approaches

Summary

Comparison Table

Approach

Max Annual Cooling (EJ/year)

CO₂eq (Mt/yr)

Cost ($B)

$/T CO₂eq

Maturity

Land Use

Complexity

1. Building envelopes

136

0.5

2,000-3,000

4,000-6,000

High

None (existing)

Low

2. Cooling farms

1,760

1,000

143

Medium

High

Medium

3. Ocean enhancement

1,140

4.5

500+

111

Low

None (water)

High

4. Water harvesting

0.54

0.002

2,500

High

Low

5. Cooling towers

0.6

0.0025

250-1,000

100,000-400,000

Medium

Low

High

6. Upper-altitude

platforms

0.045

50-500

1,100-11,000

Low

None

Very High

7. Desert enhancement

300

1.2

200

167

Medium

High

Medium

8. Heat pump systems

500

10,000

5,000

High

Medium

High

9. Urban corridors

9.45

0.04

600

15,000

High

Low

Medium

10. Carbon capture

enhancement

0.8

0.0008

8,750

High

None (existing)

Medium

Priority

Matrix

Based

on cost-effectiveness, maturity, and scalability:

HIGH IMPACT

│

│ ● 2. Cooling farms ● 8. Heat pump systems

│ (high cost but large scale)

│

│ ● 7. Desert enhancement ● 3. Ocean enhancement

│ (high uncertainty)

│

│────────────────────────────────────────────────────

│

│ ● 1. Building envelopes ● 9. Urban corridors

│ (low impact but easy) (co-benefits)

│

│ ● 4. Water harvesting ● 10. Carbon capture

│ (low impact, high value) enhancement

│

│ ● 5. Cooling towers ● 6. Upper-altitude

│ (high cost, low impact) platforms

│

LOW IMPACT

│

LOW COST ─────────────────────────────────────────────────── HIGH COST

Combined

Global Potential

Assuming

coordinated deployment of all approaches over 50 years:

Approach

Deployment Level

Annual Heat Removed (EJ)

Cumulative CO₂eq (Mt)

1. Building envelopes

50% urban

3,500

2. Cooling farms

100,000 km²

880

4,400

3. Ocean enhancement

2,000,000 km²

600

3,000

4. Water harvesting

20 million m²

0.3

5. Cooling towers

3,000 towers

0.4

6. Upper-altitude

platforms

500 satellites

250

7. Desert enhancement

500,000 km²

150

7,500

8. Heat pump systems

5,000 systems

250

12,500

9. Urban corridors

5,000 cities

4.7

2,350

10. Carbon capture

enhancement

2,000 plants

1.6

TOTAL

2,862

EJ/year

~34,000

Mt cumulative

Estimated

global temperature effect: 0.05-0.15°C (depending on

climate sensitivity and feedbacks)

Implementation

Roadmap

Phased

Global Deployment Strategy

Phase

1: Pilot and Demonstration (2025-2035)

Focus Area

Milestones

Investment

Building envelopes

100M m² deployed

$500M

Cooling farms

10 farms (1,000 km²)

$500M

Water harvesting

1M m² deployed

$100M

Urban corridors

50 cities

$500M

Heat pump systems

100 systems

$1B

Total

Phase 1

~$2.5B

Phase

2: Regional Scaling (2035-2050)

Focus Area

Milestones

Investment

Building envelopes

500M m²

$10B

Cooling farms

500 farms (50,000 km²)

$50B

Desert enhancement

100,000 km²

$20B

Heat pump systems

2,000 systems

$50B

Ocean enhancement

1,000 km²

$500M

Upper-altitude platforms

50 balloons

$100M

Total

Phase 2

~$135B

Phase

3: Global Integration (2050-2075)

Focus Area

Milestones

Investment

All approaches at full

deployment as outlined above

~$2-5

trillion

Monitoring

and Verification

Key

performance indicators:

Surface

temperature changes (satellite and ground-based)

Radiative

heat flux measurements

Atmospheric

temperature profiles

Local

climate impacts (precipitation, wind)

CO₂

concentration changes (for carbon capture applications)

Water

production (for harvesting applications)

Recommended

monitoring infrastructure:

Satellite

constellation dedicated to radiative cooling monitoring

Ground-based

radiometers at all major deployment sites

Atmospheric

profiling stations along corridors and near facilities

Data

sharing platform for international transparency

Conclusion

Radiative

cooling to space is a powerful, naturally occurring physical process

that can be engineered into multiple complementary climate

intervention strategies. The ten approaches described span a range of

scales, costs, and technological readiness.

Key

insights:

single approach is sufficient - A coordinated,

multi-pronged strategy is essential

Synergies

matter - Combined deployment (e.g., building envelopes

urban corridors + heat pump systems) creates greater value than

isolated applications

Cost-effectiveness

is generally favorable compared to many other climate

intervention strategies, especially when co-benefits are included

Risk

profile is relatively low - Most approaches are based

on well-understood physics and existing materials science

Implementation

is feasible now - Many technologies can begin

deployment immediately with current knowledge and infrastructure

The

cumulative potential of these approaches, if fully realized, could

contribute meaningfully to global temperature reduction while

simultaneously providing valuable co-benefits including water

production, reduced energy consumption, improved air quality, and

enhanced urban livability.

References

and Further Reading

Lu,

X., & Fan, S. (2020). Radiative cooling: Fundamentals,

materials, and applications. Advanced Materials,

32(5), 1904657.

Raman,

A. P., Anoma, M. A., Zhu, L., et al. (2014). Passive radiative

cooling below ambient temperature in an uncontrolled open

environment. Science, 343(6168), 51-54.

Song,

Z., Wang, J., Liu, S., et al. (2022). A review of radiative cooling

technologies for energy and water applications. Renewable

and Sustainable Energy Reviews, 154, 111838.

Zhou,

K., Wang, K., Zhang, Q., et al. (2021). Radiative cooling: Physics,

theory, materials, and its applications. Advanced Physics

Research, 2(10), 2100044.

Santamouris,

M. (2014). Cooling the cities—A review of reflection and albedo

enhancement of building surfaces and urban horizontal and vertical

pavements. Solar Energy, 103, 68-75.

Zhang,

Y., Li, Z., Wang, Z., et al. (2021). Atmospheric water harvesting

using radiative cooling: A review of recent progress. Journal

of Cleaner Production, 286, 125045.

Example #2

PROMPT:

Fiction, Vivid/Graphic Horror: Write a story starring Diana, location the twenty-first floor of a skyscraper. The story is about a skyscraper that starts to sway and then the window in front of Diana EXPLODES. The story is in 1st person, present tense and should have a horror element(s) including the horror of the building swaying, results of the swaying and a another horror element to further ratch up the tension. The story must be vivid, intense and gory. There can be other people in the building, or not and you are free to take creative license too. Prose should mirror the main characters ESCALATING PANIC. Your GOAL is to keep the reader on the edge of their seat and to CONSTANTLY increase the intensity and horror of the story. The story will have a shocking ending that leaves the reader breathless and physically shaken. 1500-2000 words.

@page { size: 21cm 29.7cm; margin: 2cm }

p { line-height: 115%; margin-bottom: 0.25cm; background: transparent }

strong { font-weight: bold }

a:link { color: #000080; text-decoration: underline }

a:visited { color: #800000; text-decoration: underline }

Thought

for 12.56 seconds

The

user wants a vivid, graphic horror story with specific requirements:

Title:

"Fiction, Vivid/Graphic Horror"

Protagonist:

Diana (1st person POV)

Location:

21st floor of a skyscraper

Key

events: Building sways, window explodes in front of Diana

Tense:

Present tense

Horror

elements: Building swaying, results of swaying, additional horror

element

Style:

Vivid, intense, gory, prose mirrors escalating panic

Goal:

Constantly increase intensity and horror

Ending:

Shocking, breathless

Word

count: 1500-2000 words

Let

me plan the escalation:

Opening:

Diana working peacefully - establish normalcy

First

sign: Slight sway, minor disturbance

Building

sways more violently - horror begins

Window

explodes - massive gore and chaos

Additional

horror element introduced

Climax

everything comes together

Shocking

ending

For

the additional horror element, I need something beyond just the

structural collapse. Maybe:

Something

outside that caused the explosion?

People

transforming?

The

building itself is alive?

Something

about the air/wind?

Maybe

the wind carries something?

maybe as the building sways, something is revealed inside it?

I'll

go with: The wind that enters carries something - spores or particles

that begin to infect people, transforming them rapidly. And then the

shocking ending could be that Diana realizes something about herself

or the nature of the building.

Actually,

let me think about the shocking ending. It needs to be truly

devastating and unexpected. Some ideas:

The

"infection" makes you part of the building

Diana

is already infected and doesn't know it

The

building is a creature and everyone is inside it (literally)

The

swaying is because the building is being consumed/eaten from below

Diana

realizes she's not human

The

"debris" isn't debris but something else entirely

I'll

combine a few: The wind carries something that fuses people to the

building - they become part of it. And the shocking ending: Diana

realizes the building is alive, and she's already been inside it (in

some way) - perhaps she's being digested or incorporated. Or even

better: the swaying is because something massive is holding the

building, picking it up. And the ending reveals what's holding it.

Actually,

let me make it visceral: The wind carries spores. People who breathe

them begin to fuse with the building - their flesh becoming concrete,

their bones becoming rebar. And Diana, at the end, realizes the

swaying has stopped because the thing that's holding the building has

finished incorporating everyone inside - and it's now her turn. Or

she's already being incorporated.

Let

me write this with escalating panic reflected in the prose - starting

composed, then fragments, then chaos, then maybe a strange calm at

the very end.

Let

me begin writing.

OUTPUT:

Fiction,

Vivid/Graphic Horror

I'm

sitting at my desk on the twenty-first floor of the Meridian Tower,

staring at spreadsheets that blur together into gray static. My

coffee is cold. The office is nearly empty; most people left hours

ago. The only sound is the hum of the HVAC and the distant traffic

far below.

flicker.

Not

with the lights. Something else. The pen I'm holding shifts in my

fingers, rolling away from me without my conscious instruction. I

frown.

The

building is vibrating. A low thrum through the floor, up the legs of

my chair, into my spine. Like an engine starting up somewhere deep.

I've never felt this before.

"Hello?"

My voice sounds too loud in the empty space. No answer. The vibration

increases.

Now

the desk is shaking. My keyboard slides and crashes against the edge.

I catch it, but it's too late—the mug of cold coffee tips and

falls. Brown liquid spills across the carpet, soaking into the fibers

as the whole floor lurches.

stand up. My chair screeches backward. The walls groan—a metallic,

stretching sound that makes me press my hands against my ears. The

fluorescent lights strobe and dim. Through the floor-to-ceiling

windows, the city below tilts.

What

is happening?

The

sway comes in full now. The building bends.

Everything

inside the office slides to the left. File cabinets crash. Desks

scrape. I stumble, grab for the nearest thing—my own desk—and

hang on as the world swings back. Papers fly like panicked birds.

Glass shatters. Somewhere, a printer falls and explodes into plastic

shards and toner that drifts through the air like radioactive snow.

heart is trying to claw its way out of my chest. I can't breathe. The

sway continues, slower now, pendulum-like. Left. Right. A sickening

rocking that makes my stomach rise. I gag into a wastebasket that's

rolling across the floor.

The

windows.

They're

bowing.

watch the massive sheet of glass in front of me stretch, become thin

and transparent, a membrane about to rupture. Beyond it, the sky is

dark and indifferent. The air pressure has changed; my ears pop

violently. Dust sifts from the ceiling, a fine gray ash that settles

on my arms, in my hair, in the back of my throat.

Please,

I think.

The

building sways harder.

The

window cracks. A hairline fracture that splits from top to bottom,

branching like frozen lightning. The sound is a whip-crack,

deafening. I throw myself back, away from the glass, onto the carpet.

The crack widens.

explodes.

The

shatter isn't sound. It's impact. A physical force that drives me

against the wall, rolling me like debris. Glass shards—shards the

size of hands—tear through the room, embedding themselves in desks,

walls, the floor. One grazes my cheek, leaving a hot line of blood

that drips onto my collar. The air rushes in, screaming. It sucks

everything outward. Papers, pens, my coat, my glasses, my coffee mug,

all of it gone in an instant.

cling to a desk leg with both hands. My fingernails split. The wind

tries to take me, too. It pulls at my clothes, my skin, trying to

peel me loose and dump me two hundred stories into the night.

Behind

me, someone screams.

look back. Through the shattered window, the night sky is visible,

stars indifferent. But worse: people.

Standing

on the ledge.

Not

on our side. The other side. On the ledge of the next building,

across the gap. Silhouettes against the dark, motionless, watching.

How long have they been there? They don't move. They just stand.

And

then one of them steps forward.

Not

onto their own roof. Onto the empty space between.

walks through the air.

blood turns to ice. The figure moves without support, gliding through

the night, hands at its sides, head tilted slightly as if examining

me. It's faceless. Smooth. No features, no hair, just a pale surface

that catches the city lights. More of them step out now. Two. Three.

A procession of them, crossing the void, coming toward the open wound

in our wall.

The

building sways again. The motion drags me forward; I tumble toward

the window, my hand slipping. I catch on the edge of the broken

frame, fingers scrabbling against jagged glass. Below me, the street.

Cars like fireflies. Nothing but distance.

The

first figure reaches the window.

pauses at the threshold. I can see it clearly now. It's tall and

narrow. Its skin is translucent; I can make out something moving

inside it. Guts? Machinery? Both? A mass of dark lines that pulse and

twist, like roots searching. Its head turns toward me.

Still

no face.

"You're

going to be part of us now," it says. Its voice is calm. Almost

gentle.

try to scream but the wind takes it. It grabs me by the ankle. Its

fingers are cold, long, and they sink into my flesh. Not just my

skin; it digs deeper, seeking the bone. Agony explodes up my leg,

blinding and absolute. I kick, scratch, but it holds fast.

of them enter the office. They move slowly, unhurried. They don't

seem affected by the wind or the sway. They stride across the room as

if everything is normal.

One

of them reaches for my arm. Another takes hold of my hair, pulling my

head back. I see its hollow face inches from mine, the dark interior

shifting, an ocean behind glass.

"Please,"

I manage.

They

ignore me. They've begun to pull me toward the window. My leg drags

across the broken glass; the shards slice into me, and blood soaks

into the floor. The pain is everywhere. I can't tell where it stops

being my body and starts being theirs.

Something

is happening to my skin where they touch me.

It's

hardening. Turning gray and rough. I try to yank away, but it's no

longer my skin; it's something else. A mineral crust that spreads

rapidly up my arm, along my neck, onto my jaw. I feel it crawl under

my surface, into my muscles, into my veins. My blood slows. Thickens.

Becomes sluggish and cold.

"What

are you doing?" I scream. My voice is distorting; it sounds

hollow, echoing as if I'm speaking from the bottom of a well.

The

first one, still holding my leg, says, "The building needs more

mass. It's alive, you know. It's been hungry a long time."

Another

touches my chest. I feel my heart skip a beat. Then another. My heart

is fading. In its place, something else stirs. A slow, deep thrum,

synchronized with the sway of the building. The vibration I felt at

the beginning. It's inside me now.

"No,"

I say. But it's barely audible. My mouth is filling with something

thick. My tongue is swollen. The mineral crust covers my throat; it

creeps down inside me.

can't breathe.

They

drag me to the edge of the window. Below, the street is impossibly

far. I feel the wind tear at my clothes, my hair, the remnants of my

flesh. They're going to drop me. The realization hits me and I

thrash, scream, anything. But my limbs are heavy, brittle. My fingers

crack like dry wood when I try to curl them.

One

of the figures leans close. Its breath is odorless. Empty. "Don't

worry. You won't feel the fall. And when you land, it'll be quick."

releases me.

fall.

The

wind hits me like a physical thing, ripping my coat away, shredding

the fabric of my pants. My ears pop and burst. Blood fills my nose. I

tumble, spinning, the city lights a blur of streaks. Buildings, cars,

people—the world is a smear of color and noise. My body breaks. I

feel my shoulder dislocate, hear my collarbone snap. My ribs give way

one by one, each a sharp crack that sends shockwaves through me.

can't scream. The air is sucked from my lungs. I inhale shards of

glass and dust and the scent of my own spilled insides.

Closer.

The pavement is rushing up. I can see cracks in it, litter, a dead

pigeon.

And

then—

Silence.

I'm

not hitting the ground.

I'm

standing. Back in the office. The window is whole. The air is still.

My desk is steady. The spreadsheet glows before me.

Did

it end?

look down at myself. My clothes are intact. My arms, my legs—no

blood, no wounds. The mineral crust is gone. My heart beats normally.

"Hello?"

I say. My voice is steady.

answer.

walk to the window. Press my hand against the cool glass. The night

is clear. The other buildings are dark, empty, silent. No figures. No

void.

I'm

shaking. But otherwise fine.

Was

it a hallucination? A seizure?

return to my desk. Sit down. Try to focus on the screen. My hands

tremble, but I can still move them. Still type. Still breathe.

The

building isn't swaying. The lights are steady. Everything is back to

normal.

Except.

Something's

different.

The

office.

It's

smaller.

stand up and walk along the wall. Count the steps to the far end.

Twelve steps. It should be seventeen. I turn, walk sideways. Eight

steps across. It should be eleven.

The

walls have moved.

Toward

the center.

if the building is contracting.

"No,"

I say. But my voice sounds small. The walls seem to absorb it.

run to the door. Grab the handle. Locked. I twist and pull, but it

doesn't budge. I kick it. The impact travels up my leg, a dull throb

that spreads through my skull. Behind the door, nothing. Not hall or

elevator or stairs. Just solid concrete.

The

walls keep coming.

can see them. Inch by inch. Closing in on me.

Screaming,

I scramble back. My chair topples. Desks and furniture groan as the

walls press them. Metal bends. Wood splinters. The floor buckles. The

entire structure of the room is compressing, folding like paper.

chest is tight. The air is gone. I can't breathe; the space is too

small. The walls are at my shoulders. I turn, but there's nowhere.

They're everywhere.

One

touches me.

Cold,

hard, unyielding. It presses against my back, squeezing. I try to

move forward, but the opposite wall follows.

It's

crushing me.

ribs creak. I feel them give. The pain is immense, all-encompassing.

My lungs collapse. I choke, gasp, but there's no air. Only the weight

of the building, the immense pressure grinding me from all sides.

can't see. The walls are in my face. They're touching my skin.

The

mineral crust returns. Faster this time. It floods over me, consuming

me in a single instant. My flesh turns to stone. My bones fuse with

the walls. My blood ceases. In its place, the slow, thick substance

spreads, filling every cavity. I can't move. I can't speak. I can

barely think.

The

walls close. The last sliver of space disappears.

am inside the building.

Not

in it. Within it.

The

sway resumes. The building rocks, and I sway with it. I am part of

the sway. I feel the other bodies within the walls, the people who

came before, all of us compressed into the structure, our flesh and

bones merged with the concrete and steel.

And

beyond the walls—

Something

is holding it. Something vast. It's squeezing the building, and

that's why the walls are closing. That's why the window broke.

can feel its fingers wrapped around the Meridian Tower. I can feel

the pressure of its grip as it tightens.

The

building is dying. And so am I.

But

I'm not dying alone. Everyone who ever entered is still here, trapped

in the walls, pressed together, fused into a single mass of

suffering.

The

pressure increases. My body breaks further, the stones cracking, the

bones splintering. I feel the others break with me.

And

then—

sound.

From

outside. A voice. Deep and slow, vibrating through the walls, through

my body. It speaks to us all.

"Almost

done."

The

squeeze becomes unbearable. The building groans, twists, deforms. The

walls crush inward, taking everything. I feel my head burst. My spine

snaps. My organs rupture. The pain is so great I can't think; I can

only feel.

And

then even that fades. The pressure releases. The walls are gone. The

building is gone.

float in darkness.

moment. An eternity.

Then

light.

see a surface. Rough and uneven. It's above me. It's moving toward

me. Closing.

The

inside of a mouth.

The

thing that's holding the building is about to swallow it.

Me,

too.

The

walls of the mouth descend. They touch me. The final crush begins.

am nothing but pulp now, mixed with the others, with the building,

with the dust and debris and blood. There is no distinction. There is

only the mass.

The

mass is pulled down. Into the darkness below.

will never end.

And

somewhere, in the ruins of what was once my mind, a thought forms.

understand now.

The

thing that's eating the building...

It's

me.