DeepSeek unveils an experimental version of its V4 Flash model that can understand visual prompts, saying it nears the performance of Anthropic's Opus 4.8 (Bloomberg)
DeepSeek says its new vision-capable V4 Flash experiment is close to Anthropic’s Opus 4.8 on multimodal agent tests.
The model is live on DeepSeek’s API platform as `deepseek-v4-flash-vision-exp`. DeepSeek says it keeps V4 Flash’s text abilities while adding mixed text-and-image input. Images are billed as tokens at V4 Flash pricing, with support across Chat Completions, Messages, and Responses. Techmeme's note
The model is live on DeepSeek’s API platform as `deepseek-v4-flash-vision-exp`. DeepSeek says it keeps V4 Flash’s text abilities while adding mixed text-and-image input. Images are billed as tokens at V4 Flash pricing, with support across Chat Completions, Messages, and Responses. Techmeme's note
score 8