DeepSeek-V4-Flash-Vision-Exp
#13 en Modèles multimodauxDeepSeek · v4 · flash vision exp · 7× · vu le 24 août 2026
DeepSeek-V4-Flash-Vision-Exp is an experimental multimodal model from DeepSeek, released on August 21, 2026 on the DeepSeek API platform. It extends DeepSeek-V4-Flash with image understanding (text and image input, text output), while maintaining pure text capabilities (agents, Reasoning, world knowledge) at the V4-Flash level. According to the manufacturer, the model demonstrates significant progress on visual agent Benchmarks compared to V4-Flash and approaches the performance of Anthropic's Opus 4.8; however, these figures come from self-conducted, not independently verified tests. The model is available via API under the identifier "deepseek-v4-flash-vision-exp" and supports image formats including JPEG, PNG, GIF, and WebP.