DeepSeek ships V4-Flash Vision Exp multimodal API with free Files reuse
DeepSeek’s API changelog (Aug 21) launches experimental DeepSeek-V4-Flash-Vision-Exp (`deepseek-v4-flash-vision-exp`) for multimodal vision: JPEG/PNG/GIF/WebP via base64, public URL, or Files API `file_id`, on Chat Completions, Anthropic-compatible Messages, and Responses. Text agent/reasoning scores match official V4-Flash (e.g., Terminal Bench 2.1 83.9, DeepSWE 59.3, Agents’ Last Exam 27.3, Chartography 64.3, ZeroBench Pass@5 35.0); DeepSeek says multimodal-agent results jump toward Claude Opus 4.8. Images bill at ≤384 tokens each at V4-Flash rates; Files API reuse is free. Distinct from DeepSeek-V4-Pro GA and the July 31 V4-Flash text API.





