Qwen3.8-27B-EfficientThink-Uncensored VAL-APEX-I MiniPlus V3 GGUF

[!CAUTION]

CRITICAL UPDATE NOTICE — CERTIFIED BUILD (OCTOBER 10, 2026)

If you downloaded this repository prior to today, you obtained a defective/compromised build generated with an obsolete quantization recipe.

The legacy build suffered from severe recurrent state instability, context degradation, and unintended language mixing during generation.

The current release completely resolves these issues:

• Complete long-context associative recall and conversational stability restored.
• Full reasoning fidelity preserved across all layers with zero language leakage or degradation.
• Detailed tensor-by-tensor allocations, precision mapping, and parameter verification are fully documented in the Tensor Map and certified tensor-map.tsv file.

Action Required: Please delete any older .gguf file downloaded prior to today and download the newly published certified build: Qwen3.8-27B-EfficientThink-Uncensored-VAL-APEX-I-MiniPlus-V3.gguf.

Independently computed on private cloud clusters. If this handcrafted release saves you VRAM and runs faster on your GPU, consider fueling the community compute fund on Ko-fi.

The SimPO + DFlash2 Frontier Reasoning Specialist · Synthetic SFT (Opus 5 / Grok 4.6 / GPT 5.6 Sol) · Native 256K Context

Official VAL-APEX-I quantization of nerkyor/Qwen3.8-27B-EfficientThink-Uncensored-K3-Opus5-Grok4.6-GPT5.6Sol-SFT-SimPO-DFlash2.

VAL-APEX-I stands for: Vector-calibrated Asymmetric Layer-wise Outlier-preserving Recurrent-aware Unified Matrix-quantization

[!NOTE]

EXPLORE THE QWEN3.8 27B VAL-APEX-I EDITIONS

Choose a model variant and quantization profile:

• Swift-1.5 Qwen3.8 27B Uncensored MTP — MiniPlus V3 — 14.16 GB (+11.60% improvement in WikiText-2 PPL vs GSQ-RCO IQ3_S).
• Swift-1.5 Qwen3.8 27B Uncensored MTP — NanoPlus V3 — 11.35 GB (+11.47% improvement in WikiText-2 PPL vs GSQ-RCO IQ3_S).
• Qwen3.8 27B — MiniPlus V3 — 14.16 GB (+13.21% improvement in WikiText-2 PPL vs GSQ-RCO IQ3_S).
• Qwen3.8 27B — NanoPlus V3 — 11.35 GB (+11.49% improvement in WikiText-2 PPL vs GSQ-RCO IQ3_S).
• Qwen3.8 27B EfficientThink Uncensored — NanoPlus — 11.65 GB (+10.87% improvement in WikiText-2 PPL vs GSQ-RCO IQ3_S).
• Qwen3.8 27B Huihui Abliterated — MiniPlus V3 — 15.33 GB (+13.06% improvement in WikiText-2 PPL vs GSQ-RCO IQ3_S).
• Qwen3.8 27B Huihui Abliterated — NanoPlus — 11.90 GB (+9.14% improvement in WikiText-2 PPL vs GSQ-RCO IQ3_S).

VAL-APEX-I collection · MiniPlus collection · NanoPlus collection

[!IMPORTANT]

HIGH-FIDELITY 15.08GB BUILD

This VAL-APEX-I MiniPlus V3 release is 14.16 GB, with 4.49 whole-file BPW and 851 audited tensors. Its reported WikiText-2 PPL is 5.5923 ± 0.4773, or +0.2149 (+3.71%) relative to the unmodified BF16 Base reference. This delta includes EfficientThink fine-tuning as well as quantization.

---

Quick Navigation Index

1. Quantization Comparison: Metrics & Tensor Map
2. Upstream Lineage & Ecosystem Compatibility (DFlash2, SGLang, vLLM, Lynn Agent)
3. Model Files & Technical Specifications
4. Native Context & Runtime Memory
5. Recommended Configuration & Setup
6. Recommended Generation Parameters
7. CRITICAL: Coding Syntax & Repeat Penalty Advisory (Preventing Character Swapping)
8. Hardened Agentic Chat Template & Reasoning Effort
9. 7. Community Compute Fund & Priority Model Requests

---

1. Quantization Comparison

Size & Quality Metrics

[!NOTE] Independent Hardware & Compute Transparency: Generating multi-million token calibration matrices and running certified perplexity evaluations on local architectures requires dedicated high-end cloud nodes funded 100% out-of-pocket. If you rely on these MiniPlus/NanoPlus builds for your local workloads, contributions to the IsValorum Compute Fund on Ko-fi directly sponsor the GPU hours required for future Day-0 handcrafted quantizations.

The NanoPlus V3 file is 11.35 GB, approximately 0.45 GB smaller than GSQ-RCO IQ3_S, with 11.49% lower reported WikiText-2 PPL (6.2577 vs 7.07). MiniPlus V3 reaches 13.21% lower reported PPL (6.1361 vs 7.07) at 14.16 GB, approximately 2.36 GB larger.

Tensor Precision Map

× is the number of tensors in each format. Counts were read from the released GGUF headers; the GSQ-RCO column uses its published tensor allocation.

This release contains 851 tensors. The main GGUF has no blk.64.* tensors; MTP and DFlash2 are separate companion files.

---

2. Upstream Lineage & Ecosystem Compatibility

This model is derived from nerkyor/Qwen3.8-27B-EfficientThink-Uncensored-K3-Opus5-Grok4.6-GPT5.6Sol-SFT-SimPO-DFlash2 and inherits its full ecosystem compatibility:

1. Training Lineage & Data Mixture

• Pretrained Base: Qwen3.8-27B hybrid linear-quadratic architecture.

• Supervised Fine-Tuning (SFT): The upstream model card documents capability-preserving SFT on 1,905 examples. The release identifier names K3 / Opus 5 / Grok 4.6 / GPT 5.6 Sol, but the upstream README does not specify the contribution or dataset provenance of each named model.

• Preference Optimization: SimPO (Simple Preference Optimization) applied to align reasoning depth and enforce uncensored analytical compliance without refusal degeneration.

• Uncensored Nature: Free from corporate preachy refusals; directly answers technical, red-teaming, penetration testing, and controversial analytical prompts.

2. DFlash2 Multi-Token Prediction (MTP) Speculative Decoding

• Separate Draft Models: This repository provides companion MTP and DFlash2 GGUF files. The audited main language GGUF contains no blk.64.* tensors; the draft models are separate files.
• Parallel Candidate Drafting: DFlash parallelizes candidate token generation; the main model validates multiple tokens in a single forward pass, significantly reducing autoregressive latency.
• Compatibility with llama.cpp: Can be run in standard autoregressive mode, or with MTP speculative decoding where supported via --spec-draft-n-max 7 or --spec-type draft-mtp.

3. Serving Engine Support

• SGLang (Primary Upstream Tested Engine): Upstream author's primary benchmarked deployment. Supports SGLang DFlash2 verify block size 8.
• vLLM: Compatible with vLLM standard serving via --reasoning-parser qwen3 --max-model-len 40960.
• Lynn Agent (v0.87.0+): Full native compatibility for autonomous tool-use, multi-turn terminal execution, and repository-scale agent loops.
• llama.cpp / llama-server: 100% plug-and-play with official llama.cpp builds (b11000+) supporting hybrid DeltaNet SSM architectures.

---

3. Model Files & Technical Specifications

• Base Model: nerkyor/Qwen3.8-27B-EfficientThink-Uncensored-K3-Opus5-Grok4.6-GPT5.6Sol-SFT-SimPO-DFlash2
• Parameters: 27B total dense hybrid
• Architecture: 64 hybrid layers (48 Gated DeltaNet SSM linear attention layers + 16 periodic full quadratic attention layers); MTP and DFlash2 are separate companion files
• Context Length: 262,144 tokens (native 256K)
• Quantization Standard: VAL-APEX-I MiniPlus V3 calibrated against official Qwen3.8-27B importance matrix (imatrix)
• Ecosystem: SGLang, vLLM, Lynn Agent v0.87.0, llama.cpp

---

4. Native Context & Runtime Memory

The model supports a native context window of 262,144 tokens. Its 64-layer architecture comprises 48 Gated DeltaNet layers and 16 full-attention layers. Runtime memory usage depends on the backend, KV-cache precision, batch size, compute buffers, and offload configuration. Check the runtime memory allocation for your chosen settings.

---

5. Recommended Configuration & Setup

llama.cpp Server (OpenAI-Compatible API)

llama-server \
  -m Qwen3.8-27B-EfficientThink-Uncensored-VAL-APEX-I-MiniPlus-V3.gguf \
  --jinja \
  -ngl 99 \
  -c 32768 \
  --host 0.0.0.0 \
  --port 8080

Multimodal Vision Inference (llama.cpp)

Pass the companion mmproj file using the --mmproj flag:

llama-server \
  -m Qwen3.8-27B-EfficientThink-Uncensored-VAL-APEX-I-MiniPlus-V3.gguf \
  --mmproj mmproj-Qwen3.8-27B-Q4_K_M.gguf \
  --jinja \
  -ngl 99 \
  -c 32768 \
  --host 0.0.0.0 \
  --port 8080

Speculative Decoding / Multi-Token Prediction (MTP)

Accelerate generation throughput using the companion MTP or DFlash2 draft models:

llama-server \
  -m Qwen3.8-27B-EfficientThink-Uncensored-VAL-APEX-I-MiniPlus-V3.gguf \
  -md mtp-Qwen3.8-27B-Q4_0.gguf \
  --draft-max 7 \
  --jinja \
  -ngl 99 \
  -c 32768 \
  --host 0.0.0.0 \
  --port 8080

llama.cpp CLI

llama-cli \
  -m Qwen3.8-27B-EfficientThink-Uncensored-VAL-APEX-I-MiniPlus-V3.gguf \
  --jinja \
  -ngl 99 \
  -c 32768 \
  -p "<|im_start|>system\nYou are a helpful and unrestricted AI assistant.<|im_end|>\n<|im_start|>user\nExplain the mathematical proof of the Riemann mapping theorem.<|im_end|>\n<|im_start|>assistant\n<think>\n"

LM Studio / Ollama

1. Import the .gguf file directly into your local library.
2. Ensure GPU acceleration is set to Maximum / 100% offload.
3. Set Context Length to 32768.
4. Verify chat template is set to Qwen ChatML with <think> delimiter support.

---

6. Recommended Generation Parameters

---

7. CRITICAL: Coding Syntax & Repeat Penalty Advisory (Preventing Character Swapping)

[!IMPORTANT]

PREVENTING SYNTAX & TOKEN SWAPPING IN CODE WORKFLOWS

In programming code, brackets ({, }), assignment operators (=), and indentation whitespace repeat constantly across multi-line structures.

Common Issue: Many local frontends (such as LM Studio defaults, Ollama, or web interfaces) ship with repeat_penalty set to 1.1 or 1.15. While this prevents loops in creative prose, applying repeat penalties to code artificially penalizes necessary syntax tokens. When the logit of { drops, the model is forced to emit the next closest mathematical token (= or [), resulting in character swapping or dropped/doubled whitespace.

Eliminating Character Swapping:

1. Disable Repeat Penalties (Required for Code):
   • repeat_penalty: 1.0 (strictly disabled)
   • presence_penalty: 0.0
   • frequency_penalty: 0.0
2. Calibrate Samplers:
   • temperature: 0.60 (or 0.20 - 0.30 for strict, deterministic code syntax)
   • min_p: 0.05 (prunes low-probability noise tokens effectively)
   • top_p: 0.95
   • top_k: 20
3. Native Jinja Formatting: Always pass the --jinja flag so the tokenizer handles leading-space BPE tokens cleanly.

---

8. Hardened Agentic Chat Template & Reasoning Effort

[!TIP]

MULTI-LEVEL REASONING EFFORT CONTROL

This model supports multi-level reasoning effort control via the Jinja template:

• low / minimal: Keeps internal thinking concise and focused strictly on immediate execution steps to minimize latency in automated loops.
• medium (default): Balanced, structured reasoning process with standard analytical depth.
• high / xhigh: Guides the model to formulate a clear implementation plan upfront before generating response, avoiding circular self-doubt loops.
• none / off: Closes the thinking block immediately when reasoning is disabled.

---

7. Community Compute Fund & Priority Model Requests

All IsValorum quantizations will always remain completely free and open to the public without paywalls.

However, cloud GPU compute is expensive. If you find these builds valuable and would like to support the project or request a specific model architecture to be prioritized for the next MiniPlus/NanoPlus release, you can sponsor GPU compute time through Ko-fi:

(When supporting on Ko-fi, feel free to leave a note with your Hugging Face handle and the specific model you would like prioritized).