DeepSeek
DeepSeek V4 Flash 0731 (Reasoning, Max Effort)
DeepSeek V4 Flash is an efficiency-optimized Mixture-of-Experts model from DeepSeek with 284B total parameters and 13B activated parameters, supporting a 1M-token context window. It is designed for fast inference and...
DeepSeek release notesModel Snapshot
TypeLanguage
ReleaseApr 23, 2026
Context1,048,576
Tagsreasoning, tool-use, implicit-caching
Overall
Rank #35 across 600
49.9
Coding
Rank #39 across 266
69.1
Blended Price
Rank #50 across 379
$0.175/M
Strength Profile
Percentile score by analysis domain.
* Cost is inverted: lower input, output, and blended prices rank higher.
Benchmark Percentiles
Higher bars mean stronger relative placement.
Model Details
Runtime
Max Output65,536
Streaming speed is not measured for this model yet.
Pricing
Blended$0.175/M
Input$0.140/M
Output$0.280/M
Catalog$0.130/M in / $0.260/M out
Metadata
Modalities
All Benchmarks
| Metric | Domain | Value | Rank |
|---|---|---|---|
| Artificial Analysis Intelligence Index | overall | 49.9 | #35 |
| Artificial Analysis Coding Index | coding | 69.1 | #39 |
| GPQA | reasoning | 90.8% | #31 |
| Humanity's Last Exam | reasoning | 36.8% | #35 |
| SciCode |