DeepSeek released V4-Flash-Vision-Exp, an experimental multimodal model that adds image understanding to its V4-Flash text model.

Source Brief

DeepSeek released V4-Flash-Vision-Exp, an experimental multimodal model that adds image understanding to its V4-Flash text model. On DeepSeek's multimodal agent benchmarks, it approaches Anthropic's Opus 4.8 and sometimes beats it.