DeepSeek V4 Flash
DeepSeek V4 Flash (0731) is DeepSeek's efficiency model, released July 31, 2026, at $0.14 per million input tokens and $0.28 per million output — the cheapest well-known model tracked by Artificial Analysis. It scores 52 on the Intelligence Index, matching GPT-5.6 Luna and Gemini 3.6 Flash at a fraction of their prices.
Spec sheet
| Maker | DeepSeek |
|---|---|
| Released | 2026-07-31 |
| API ID | deepseek-v4-flash |
| Parameters | Not disclosed |
| Context window | 1M tokens |
| Modalities in | text |
| Modalities out | text |
| API availability | DeepSeek API · DeepInfra · SiliconFlow · OpenRouter (unverified) |
| Weights | Open (Open weights) |
| Pricing (per MTok) | $0.14 in / $0.28 outas of · source |
Benchmarks
| Benchmark | Score | Measured | Source |
|---|---|---|---|
| AA Intelligence Index | 52 | artificialanalysis.ai |
Dashed values are vendor-published and not yet reproduced in our harness — methodology.
Changelog
- DeepSeek V4 Flash 0731 released at $0.14/$0.28 per MTok — the frontier price floorsource
Head-to-head
Related news
Profile published August 11, 2026; data last changed August 11, 2026. Every value carries its own measurement or verification date. Flagged values are best-known public figures pending confirmation.