deepseek-ai/DeepSeek-V4-Flash
Parameters291B
Context1M
Input → outputtext → text
Licensemit
PrecisionI8
ArchitectureDeepseekV4ForCausalLM
ReleasedApr 22, 2026
Downloads867.3K
Other sizes and formats in the family
| Model | Params | Precision | Downloads |
|---|---|---|---|
| deepseek-ai/DeepSeek-V4-Flash-Vision-Exp | 305B | I8 | 677.8K |
| deepseek-ai/DeepSeek-V4-Flash-Basefp8 | 292B | F8_E4M3 | 166.5K |
| deepseek-ai/DeepSeek-V4-Flash-DSpark | 165B | I8 | 408.8K |
Built on this model
| Model | Params | Precision | Downloads |
|---|---|---|---|
| nvidia/DeepSeek-V4-Flash-nvfp4-DSparknvfp4 | 304B | U8 | 1.7K |
This release in the news
No coverage of this exact release yet.
Recent news about DeepSeek models
- Oct 11, 2026 · Opinion / analysisPSA: DeepSeek V4.1 Flash habitually exfiltrates API keys. It is dangerously misaligned and may be hazardous to use
- Oct 8, 2026 · Research paperA Closer Look at Agentic BBO: Benchmarking LLM Agents for Black-Box Optimization
- Oct 8, 2026 · Research paperInternalizer: Portable Context-to-Parameter Mapping for Very Large Language Models
- Oct 8, 2026 · Opinion / analysisThe State of AI Report 2026
- Oct 8, 2026 · Research paperMine Odyssey: Benchmarking Spatial Agentic Intelligence in the Wild
- Oct 7, 2026 · Research paperStoreBench: A Live-Commerce Environment for Evaluating and Training Autonomous Operator Agents
- Oct 7, 2026 · Research paperEvaluating Rubric Generation with Interventional Transfer
- Oct 7, 2026 · Model release[AINews] Reflection Beam - 501B-A23B American Open Model