DeepSeek V4 Flash High Preview
Key facts · verified Oct 29, 2026
DeepSeek V4 Flash High Preview is a DeepSeek model. It was retired on 10 September 2026. Its last listed API price was $0.44 per million input tokens and $1.32 per million output tokens (third-party listing via LMArena, not a DeepSeek price; model retired 10 September 2026). Its context window is 1M tokens. It ranks #99 on LMArena's text leaderboard (Elo 1439, read 1 Nov 2026). Source: api-docs.deepseek.com.
DeepSeek V4 Flash High Preview is a preview version of DeepSeek's V4-Flash model, with open weights under the MIT licence and a 1M-token context window. DeepSeek retired V4 Flash on 10 September 2026; requests to deepseek-v4-flash are now served by DeepSeek V4.1-Flash. On LMArena's text leaderboard it holds 1439 Elo (rank 99, high effort).
Full Specifications
$0.44 / $1.32 per million tokens | |
|---|---|
| Identity | |
| Developer | DeepSeek |
| Retired | 10 September 2026 |
| Status | Retired. DeepSeek retired V4 Flash on 10 September 2026; requests to deepseek-v4-flash are now served by DeepSeek V4.1-Flash. |
| Licence | MIT |
| Self-hostable | Yes |
| Cost | |
| Blended $/1M tokensinput × 0.75 + output × 0.25 | $0.660 / 1M tokens |
| Input price | $0.440 / 1M tokens |
| Output price | $1.320 / 1M tokens |
| Capacity | |
| Context window | 1M tokens |
| Measured quality | |
| LMArena Elo | 1439 (checked Nov 2026) |
| LMArena rank | Rank 99 |
| Elo per dollarLMArena Elo ÷ blended $/1M tokens | 2180 |