DeepSeek DSpark
Q8_0 · speculative companion · Runtime companion
The selected Q8_0 DSpark companion from Unsloth's DeepSeek V4.1 Flash 1017 distribution, retained for the tested native DeepSeek runtime.
Benchmark results · Files · Download · Runtime · Provenance · Full collection
Join Infinity on Discord
Get model support, request models and quants, discuss benchmarks, and follow new releases and model improvements.
All model tests, per-question scores, experiment history and scoring methods
Qwen effort study — separate full-404 cohort: medium 82.87/100 versus xhigh 84.41/100; paired difference +1.54 points (95% interval [-1.71, +4.82]). 404 collected questions, 403 valid questions graded per effort; one invalid reference annotation excluded symmetrically. The historical 202-question cross-model comparison below is unchanged. Study, timing and blind subjective results.
Preserved and documented by isilicon / Mango. The original model, modification and quantization remain the upstream creators' work. This repository adds attributed preservation and exact-configuration evaluation, with the selected source bytes unchanged.
At a glance
| Property | Recorded value |
|---|---|
| Artifact | GGUF dependency |
| Selected source payload | 7.98 GB / 7.43 GiB |
| Primary selected files | 1 |
| Evaluation status | Runtime companion |
| Source publisher | unsloth |
| Source revision | 85f81fb970add013033f25e94006713ca2c0aaf2 |
| Declared license | mit |
GB is decimal and GiB is binary. Payload size is disk/storage size, not a claim that the model fits in the same amount of VRAM. Runtime, KV cache, working memory and companions add their own requirements.
Mango evaluation
No standalone Mango quality score has been published for these exact files. This is a companion artifact; the consuming model's result belongs to the full target/runtime configuration.
Related configurations measured separately
These results describe the linked configurations, not this artifact. A change of quant, modification lineage, draft, engine or tokenizer can change the result.
| Different configuration | Its weighted score | Its workload min |
|---|---|---|
| DeepSeek V4.1 Flash · GSQ–RCO IQ2_XS · AtomicChat | 83.19 | 71.43 |
A completed score requires a native preflight, matching frozen prompts, 202 suite answers, 12 agent cases, archive verification and the same reference graders. The benchmark methodology describes the protocol; publisher-side results are retained separately below.
Runtime and compatibility
Use only with a compatible DeepSeek target and native speculative path. Target/draft acceptance and end-to-end speed must be measured as a configuration; the parent's score cannot be assigned to the draft alone.
Auxiliary file; no standalone Mango benchmark or independent model score.
Companion artifacts
| Consuming model | Connection |
|---|---|
| DeepSeek V4.1 Flash · GSQ–RCO IQ2_XS · AtomicChat | Use the compatible parent runtime and exact target revision |
Files and integrity
| Selected file · original path | GB | Identity prefix |
|---|---|---|
| dspark-DeepSeek-V4.1-Flash-1017-Q8_0.gguf | 7.97691 | 6ba6ac1b331c5406… (SHA-256) |
Full paths, byte sizes and complete SHA-256/Git blob identities are in SOURCE_MANIFEST.json. Available original README/license/notice files are also preserved. MIRROR_STATUS.json records the earlier source-file verification; MODEL_CARD_STATUS.json records this documentation refresh separately.
Download and use
The commands pin the already-verified mirror weight revision. Documentation can advance on main without changing that preserved weight snapshot. Keep original subfolders and all shards.
python -m pip install -U huggingface_hub
hf download isilicon/mango-deepseek41-dspark-q8 \
dspark-DeepSeek-V4.1-Flash-1017-Q8_0.gguf \
--revision f0721c818cc1716d63bf9d183af67c3e27080f16 \
--local-dir ./weights/deepseek41-dspark-q8
Load example and setup scope
This component has no independent chat launch command. Install it through the consuming model's compatible runtime and companion layout.
Provenance and license
Only the named dspark file is selected here. The original Unsloth card describes the broader DeepSeek family and other quants; those full-model instructions are not a claim that this companion is independently usable.
| Responsibility / identity | Attribution |
|---|---|
| Mirror / local evaluation | isilicon · Mango |
| Immediate source | unsloth/DeepSeek-V4.1-Flash-1017-GGUF |
| Pinned source commit | 85f81fb970add013033f25e94006713ca2c0aaf2 |
| Declared upstream base(s) | deepseek-ai/DeepSeek-V4.1-Flash-1017 |
| Source-declared license | mit |
Original notices, base-model conditions and the publisher's stated limitations still apply. Downloadable preservation does not add new license rights. Source-described refusal modification and publisher-side benchmark claims have not been independently re-measured by this Mango task cohort.
Original publisher documentation
The unchanged upstream README and pinned source card retain the original usage guidance, technical discussion, attribution, caveats and citations.
Read the complete original publisher cardSource reference. The following is the original publisher's documentation. Its other quants, full-model files, context claims and benchmark tables can exceed the selection in this mirror. Relative asset/file URLs are qualified to the pinned original source; publisher measurements remain separate from Mango measurements.
Read our How to Run DeepSeek-V4.1 Flash Guide!
See Unsloth Dynamic 3.0 GGUFs for our quantization benchmarks.
- Run on a 128GB+ RAM device for the smallest 1-bit, 3-bit on 256GB or full precision on 320GB. Read Guide
- To run, please use our llama.cpp PR or the Unsloth Desktop app.
- Get up to 3x faster inference with Unsloth vs other ngram/CPU/VRAM repos./
To swap the engram (MXFP8 to Q6_K / Q5_K / IQ4_NL, see this guide)
DeepSeek-V4.1-Flash-1017
Introduction
DeepSeek-V4.1-Flash-1017 is a minor update to DeepSeek-V4.1-Flash, delivering substantial improvements across reasoning, coding, agentic tasks, and visual benchmarks. Alongside these gains, the model reduces task execution time and output length, enabling users to complete tasks faster and at lower cost and advancing the performance–cost Pareto frontier across multiple agent harnesses.
| Benchmark | DS-V4.1-Flash-1017 | DS-V4.1-Flash | K3 | GLM-5.3 | GPT-5.6 Sol | Opus 5 |
|---|---|---|---|---|---|---|
| HLE w/o tools | 42.6 (44.0†) | 36.8 (39.1†) | 43.5 | 42.0† | 44.5 | 56.3 |
| Terminal-Bench 4.0 | 42.3 | 31.2 | 12.6 | 37.9 | 39.9 | 51.8 |
| DeepSWE v1.1 | 77.0 | 74.2 | 67.5 | 66.9 | 73.0 | 74.0 |
| ProgramBench | 29.1 | 20.3 | 17.5 | 19.0 | 23.0 | 37.0 |
| NL2Repo-Bench | 68.2 | 65.4 | 58.0 | 58.0 | 56.8 | 75.3 |
| Automation-Bench v1.0.6 | 60.5 | 54.8 | 46.7 | 48.8 | 45.8 | 50.3 |
| Agents’ Last Exam | 35.4 | 31.8 | 27.6 | 28.5 | 26.7 | 28.6 |
| Apex Bench v1.1 | 72.1 | 64.6 | 47.8 | 56.6 | 54.3 | 65.8 |
| Chartography (w/tools) | 83.0 | 78.9 | 68.1 | — | 79.9 | 84.0 |
| BabyVision (w/tools) | 92.5 | 89.6 | 85.7 | — | 88.9 | 94.1 |
| ZeroBench-main (w/tools) | 51.0 | 49.0 | 41.0 | — | 53.0 | 52.0 |
† denotes the text-only subset of HLE. We use the Claude Code harness for Terminal-Bench 4.0 and the Mini-SWE-V2 harness for DeepSWE v1.1.
Prompt Encoding
This release does not include a Jinja-format chat template. The encoding folder contains a self-contained Python reference implementation (encoding.py) with test cases for multi-turn conversations, tool calling, thinking mode, numeric reasoning effort, mid-conversation system messages, and interleaved image content.
For production use, we additionally release deepseek-recipe, a set of Rust libraries with Python bindings that provides the same prompt format as a maintained, protocol-aware toolkit. It converts Messages, Chat Completions, and Responses API requests into the Conversation format, encodes them into DeepSeek V4 and V4.1 prompts or token IDs, and parses model output back into complete or streamed responses — covering thinking, tool calls, images, and generation settings. Model inference, tool execution, and HTTP transport are left to the caller.
Minimal Inference
Please refer to the inference folder for instructions on weight conversion and running inference locally.
Recommended sampling parameters:
| Parameter | Value |
|---|---|
temperature |
1.0 |
top_p |
1.0 |
context_window |
1M tokens |
max_tokens |
≥ 256K |
License
This repository and the model weights are licensed under the MIT License.
Citation
@misc{deepseekai2026deepseekv41flash,
title={DeepSeek-V4.1-Flash: Pushing the Limits of KV Cache Compression},
author={DeepSeek-AI},
year={2026},
}
Contact
If you have any questions, please raise an issue or contact us at service@deepseek.com.
Preservation record
Selected source bytes are verified against pinned metadata; the original eight graded mirrors also passed the previously recorded anonymous 1 MiB weight-prefix checks. This is not a full independent-device restore/load test. Hugging Face repository copies share a provider and do not guarantee permanent storage.