1 week ago
7
DeepSeek has released a test version of its latest AI model with multimodal capabilities. The company said it is nearing Anthropic’s Opus 4.8 in performance on multimodal agentic tasks.
The experimental DeepSeek-V4-Flash-Vision-Exp is based on the company’s V4 Flash model but can now process visual prompts, including images and screenshots, and act on the information it interprets. DeepSeek said its performance on text-based tasks remains comparable with V4 Flash, covering areas such as reasoning, agent capabilities and general world knowledge.
According to DeepSeek, the experimental model performed close to Anthropic’s Opus 4.8 in tests focused on multimodal agentic capabilities. Those capabilities allow AI systems to carry out actions with less continuous prompting or supervision from users. DeepSeek is making the model available through its API.
The move comes amid intensifying competition between Chinese and US AI companies. Chinese models are increasingly matching the performance of leading US systems at lower prices.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.
English (US) ·