
DeepSeek-VL 7B
DeepSeek's open 7B vision-language model for real-world multimodal tasks.
About
DeepSeek-VL-7B is the 7-billion-parameter flagship of DeepSeek's first vision-language family, released on March 11, 2024, in base and chat variants (plus 1.3B siblings). Designed for real-world vision-language understanding, it handles logical diagrams, web pages, formula recognition, scientific literature, natural images, and embodied scenarios.
The open release includes weights, inference code, and a Gradio demo, with commercial use permitted under the DeepSeek model license. It was among the strongest open multimodal models of early 2024 before the DeepSeek-VL2 successor arrived.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 8/10
Scored 8/10 (verified tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 112; Trending score 80/100 from researcher assessment.