Qwen2.5-VL-3B
Compact 3B open-weight vision-language model with document grounding, released January 2025.
About
Qwen2.5-VL-3B is the smallest member of Alibaba's Qwen2.5-VL family, released January 28, 2025 alongside 7B and 72B siblings. It accepts text and image input with a 129K token input window and can parse documents, videos, and screenshots, including object grounding without image rescaling.
The 3B variant is positioned for on-device and edge deployment where the 7B and 72B models are too heavy. Licensing caveat: coverage at launch described the 3B weights as free for non-commercial uses while the 7B shipped under Apache 2.0, so commercial deployers should confirm the current license on the official model card before use.
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 114; Trending score 80/100 from researcher assessment.