Model
DeepSeek V4 Flash 0731
DeepSeek · cn · 2026-07-31 · Artificial Analysis 34.3
Level with the other mid-tier flagships at 34.3, and an order of magnitude cheaper: EUR 0.80 per million output tokens against Kimi K3's EUR 15.00, with cached input at EUR 0.08. A million-token context window and high throughput make it the model to run at volume.
Input
text
Output
text
Provider listings
Text
Language and vision-language generation.
In €Out €
No models match these filters.