Skip to main content
DE

DeepSeek V4 Flash 0731

deepseek/deepseek-v4-flash-0731

DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total. This re-post-trained revision is suited for coding, reasoning, and agent workflows.

Compare
Context window
1M
Input
Text
Output
Text
Released
2026-07-31
Max output
66K
Knowledge cutoff
2025-05
Model parameters
304B
Active parameters
13B
Weights
Open
License
mit
Capabilities
Open weight, Reasoning, Structured, Temperature, Tool call

Pricing

No official pricing available.

Performance

Overall capability

52.8/100

Coding
48.4
Reasoning
52.7
Professional work
59.2
Knowledge & language
50.0
Code MigrationCodingVals AI2026-11-05
AA-LCRDeepSeek V4 Flash 0731 (Max)Knowledge & languageArtificial Analysis2026-11-05
AA-BriefcaseProfessional workArtificial Analysis2026-11-05
Humanity’s Last ExamDeepSeek V4 Flash 0731 (Max)ReasoningArtificial Analysis2026-11-05
CritPtDeepSeek V4 Flash 0731 (Max)ReasoningArtificial Analysis2026-11-05
BullshitBench v2Knowledge & languageBullshitBench2026-11-05
Terminal-Bench 4.0 (Vals rerun)CodingVals AI2026-11-05
AutomationBench-AADeepSeek V4 Flash 0731 (Max)Professional workArtificial Analysis2026-11-05

Providers

Ambientdeepseek/deepseek-v4-flash-0731Openai chat-
DeepSeekdeepseek-v4-flash-0731Custom-----
Kilo Gatewaydeepseek/deepseek-v4-flash-0731Openai chat--
NanoGPTdeepseek/deepseek-v4-flash-0731Openai chat--
OpenRouterdeepseek/deepseek-v4-flash-0731Openai chat--
TensorXdeepseek/deepseek-v4-flash-0731Openai chat--
Vercel AI Gatewaydeepseek/deepseek-v4-flash-0731Openai chat--

ModelRegistry · Catalog updated 2026-11-05