Deepseek's new vision model approaches Opus 4.8 on agent tasks

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

Why it matters: A Chinese lab closing the gap with a leading US frontier model on agent benchmarks, even using self-reported figures, adds another data point to the narrative that the US-China AI capability gap is narrowing rather than widening.