← Back to live feed1 story
The company published weights for its DeepSeek-V4-Flash-Vision-Exp model on Hugging Face and its API platform. This experimental multimodal model matches the text capabilities of DeepSeek-V4-Flash in reasoning and world knowledge and brings multimodal agent performance close to Opus-4.8.
The release creates feature parity with AI models from Moonshot and GLM. DeepSeek launched Harness 0.1.1 with out-of-the-box support for the model, and the sgl-project is developing an updated cookbook for implementation.
Sign in to suggest edits
Key sources
- SOURCE@sgl_project“bringing multimodal agent performance close to Opus-4.8”x.com
- SUPPORT@miaai_lab“DeepSeek v4 Flash Vision Exp is now open-weight!”x.com
- SUPPORT@huggingpapers“boosts agent capabilities while keeping text performance”x.com
- SUPPORT@teortaxestex“Now there's some feature parity with Moonshot and GLM”x.com
- SOURCE@vllm_project“first multimodal model in the V4 family”x.com
- SUPPORT@ilya_ai_x“experimental 305B vision model under the MIT license”x.com
- SOURCEhuggingnewshuggingnews.com