DeepSeek launches an experimental multimodal model to rival Anthropic

DeepSeek launched its experimental multimodal model V4-Flash-Vision-Exp, which can process images and text, showing performance close to Anthropic's Opus-4.8 on several benchmarks. The model surpasses Opus-4.8 on three out of eleven tests but lags on others. DeepSeek also released an updated agent harness supporting this model.