DeepSeek V4 Flash 0731
DeepSeek V4 Flash 0731 is the official open-weight release that supersedes the preview model and adds the DSpark speculative decoding module. Its 304B-parameter, approximately 13B-active MoE checkpoint uses mixed FP4/FP8 weights, supports a 1M-token context window, three reasoning-effort levels, and agentic tool use. It is downloadable from Hugging Face and is also available through the Ollama Cloud tag deepseek-v4-flash:0731-cloud.
Task Fit
Tool use, repo work, terminal workflows, and coding benchmarks.
Code generation, debugging, refactoring, and benchmark signal.
General writing, Q&A, and assistant use.
Document QA benefits from long context and instruction following.
Not marked for vision in the current library.
Not marked for image generation in the current library.
Not marked for video generation in the current library.
Not marked for voice in the current library.
Source Confidence
Variants and Quant Artifacts
Choose the artifact first; hardware fit follows from RAM, VRAM, format, and runtime.
| Quant | Format | Quality | Min RAM | Reco RAM | Runtime | Action |
|---|---|---|---|---|---|---|
| MIXED_FP4_FP8 | safetensors | balanced | 192GB | 256GB | transformers, vllm, sglang | Plan with this |
Recommended Hardware
Benchmarks
Source and Review
Similar Models
DeepSeek-R1 distilled to Llama 70B base. Strong general reasoning.
DeepSeek-R1 reasoning model distilled to 32B. Excellent at math & logic.
Alibaba's flagship Qwen3. Competitive with GPT-4 class models.