🤗 unsloth/DeepSeek-V4-Flash-0731-GGUF:DeepSeek-V4-Flash-0731 量化模型
英文标题:🤗 unsloth/DeepSeek-V4-Flash-0731-GGUF
Unsloth 发布了 DeepSeek-V4-Flash-0731 的 GGUF 量化版本,并称其 Dynamic 2.0 量化方案在准确性上优于其他主流量化版本。官方数据显示,该模型在 Terminal Bench 2.1、DeepSWE、Toolathlon-Verified 等多项 Agent 与代码任务基准上显著超过 DeepSeek-V4-Flash 预览版,并在多项测试中优于 DeepSeek-V4-Pro 预览版。开发者可使用 162GB 的 UD-Q8_K_XL 实现全精度无损运行,该版本仅比 UD-Q4_K_XL 大约 7GB;模型也已可在 Unsloth Studio 中通过 High 和 Max 推理开关运行。
4.9 KB
31.0 KB
14.9 KB
12.2 KB
422.4 KB
411.7 KB
中文
Unsloth Dynamic 2.0 可实现更高的准确性,并优于其他领先量化版本。
- 若要以全精度无损方式运行 DeepSeek-V4-Flash-0731,请使用 Q8(UD-Q8_K_XL)。该版本大小为 162GB,仅比 Q4(UD-Q4_K_XL)大 7GB。
- 有关量化分析和使用说明,请参阅我们的 DeepSeek-V4 指南。
- 现在可在 Unsloth Studio 中运行 DeepSeek-V4-Flash-0731,并可切换 High 和 Max 思考模式。

DeepSeek-V4-Flash-0731
简介
DeepSeek-V4-Flash-0731 是 DeepSeek-V4-Flash 的正式版本,取代此前的预览版本,Agent 能力得到显著增强。其模型结构与 DeepSeek-V4-Flash-DSpark 相同,即附带了一个推测解码模块。
尽管激活参数量小得多,DeepSeek-V4-Flash-0731 在以下基准测试中优于 DeepSeek-V4-Pro(预览版),并且总体上可与目前最强的闭源模型竞争。
| 基准测试 | DeepSeek-V4-Flash-0731 | DeepSeek-V4-Flash(预览版) | DeepSeek-V4-Pro(预览版) | GLM-5.2 | Opus-4.8 |
|---|---|---|---|---|---|
| Terminal Bench 2.1 | 82.7 | 61.8 | 72.1 | 81.0 | 85.0 |
| NL2Repo | 54.2 | 39.4 | 38.5 | 48.9 | 69.7 |
| Cybergym | 76.7 | 38.7 | 52.7 | - | 83.1 |
| DeepSWE | 54.4 | 7.3 | 12.8 | 46.2 | 58.0 |
| Toolathlon-Verified | 70.3 | 49.7 | 55.9 | 59.9 | 76.2 |
| Agents' Last Exam | 25.2 | 15.8 | 16.5 | 23.8 | 25.7 |
| AutomationBench Public | 25.1 | 10.8 | 12.8 | 12.9 | 27.2 |
| DSBench-FullStack † | 68.7 | 37.0 | 41.8 | 61.8 | 71.6 |
| DSBench-Hard † | 59.6 | 25.8 | 31.1 | 54.5 | 71.7 |
注释:
- 在上述公开基准中的 Code Agent 任务上,DeepSeek-V4-Flash-0731 使用即将发布的 DeepSeek Harness 最简模式作为 Agent 框架进行评测,推理强度设为
max,并使用temperature = 1.0, top_p = 0.95。 - † DSBench-FullStack 是内部全栈开发测试集;DSBench-Hard 是内部高难度代码 Agent 问题测试集。
许可证
本仓库及模型权重采用 MIT License 授权。
引用
@misc{deepseekai2026deepseekv4,
title={DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence},
author={DeepSeek-AI},
year={2026},
}
English / 原文
Unsloth Dynamic 2.0 achieves superior accuracy & outperforms other leading quants.
- To run DeepSeek-V4-Flash-0731 in full precision lossless, run Q8 (UD-Q8_K_XL), which is 162GB and only 7GB bigger than Q4 (UD-Q4_K_XL).
- See our DeepSeek-V4 guide for quantization analysis and instructions.
- You can now run DeepSeek-V4-Flash-0731 in Unsloth Studio with toggles for High and Max thinking.

DeepSeek-V4-Flash-0731
Introduction
DeepSeek-V4-Flash-0731 is the official release of DeepSeek-V4-Flash , superseding the preview version, with substantially enhanced agentic capabilities. It has the same model structure as DeepSeek-V4-Flash-DSpark, i.e. it comes with a speculative decoding module attached.
DeepSeek-V4-Flash-0731 outperforms DeepSeek-V4-Pro (Preview) on benchmarks listed below despite its far smaller activated parameter count, and is broadly competitive with the strongest proprietary models available.
| Benchmark | DeepSeek-V4-Flash-0731 | DeepSeek-V4-Flash (Preview) | DeepSeek-V4-Pro (Preview) | GLM-5.2 | Opus-4.8 |
|---|---|---|---|---|---|
| Terminal Bench 2.1 | 82.7 | 61.8 | 72.1 | 81.0 | 85.0 |
| NL2Repo | 54.2 | 39.4 | 38.5 | 48.9 | 69.7 |
| Cybergym | 76.7 | 38.7 | 52.7 | - | 83.1 |
| DeepSWE | 54.4 | 7.3 | 12.8 | 46.2 | 58.0 |
| Toolathlon-Verified | 70.3 | 49.7 | 55.9 | 59.9 | 76.2 |
| Agents' Last Exam | 25.2 | 15.8 | 16.5 | 23.8 | 25.7 |
| AutomationBench Public | 25.1 | 10.8 | 12.8 | 12.9 | 27.2 |
| DSBench-FullStack † | 68.7 | 37.0 | 41.8 | 61.8 | 71.6 |
| DSBench-Hard † | 59.6 | 25.8 | 31.1 | 54.5 | 71.7 |
Notes:
- For the Code Agent tasks among the public benchmarks above, DeepSeek-V4-Flash-0731 is evaluated with the minimal mode of DeepSeek Harness (to be released) as the agent framework, using the
maxreasoning effort level withtemperature = 1.0, top_p = 0.95. - † DSBench-FullStack is an internal full-stack development test set; DSBench-Hard is an internal test set of difficult coding-agent problems.
License
This repository and the model weights are licensed under the MIT License.
Citation
@misc{deepseekai2026deepseekv4,
title={DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence},
author={DeepSeek-AI},
year={2026},
}