DEEPSEEK TESTS VISUAL AI MODEL TO MATCH ANTHROPIC

The Chinese AI company released the test model as it continues to compete with established players in the generative AI space. The new version adds multimodal capabilities to DeepSeek's existing Flash model, enabling it to interpret and respond to image-based inputs alongside text. DeepSeek's claim positions the experimental model as a direct alternative to Anthropic's Opus 4.8, one of the leading large language models in the market. The comparison suggests the startup is closing the performance gap with Western competitors, particularly in handling complex visual reasoning tasks. Multimodal AI—systems that process both text and images—has become a competitive differentiator in the AI sector. Major models like OpenAI's GPT-4V and Google's Gemini already offer this capability, making visual understanding a standard feature rather than a premium offering. DeepSeek's push into visual AI reflects broader industry trends toward more capable, integrated AI systems. The experimental designation indicates the model remains in development and may undergo refinement before potential wider release. The announcement comes amid intensifying competition between Chinese and US AI companies. DeepSeek has gained attention for producing capable models while operating with reported lower computational costs than many Western counterparts. Anthropic, founded by former OpenAI researchers, has positioned itself as a safety-focused alternative in the generative AI market. No timeline was provided for when the experimental model might transition to a stable release or become available to users. DeepSeek typically releases its models through various channels, including API access and direct deployment options.