Qwen3-VL
by Alibaba
· QWEN3 family · released Sep 25, 2025
Open weights
Active
Open-weight vision-language family (2B/4B/8B/32B dense and 30B-A3B MoE variants) with 256K context, strong OCR/chart/document understanding, and video understanding. Positioned as an open alternative to proprietary VLM APIs.
No pricing recorded for this model yet.
Specifications
| Model type | Multimodal |
|---|---|
| Architecture | Dense |
| Parameters | Not disclosed |
| Context window | 262.1K tokens |
| Max output | Not disclosed tokens |
| Knowledge cutoff | Not disclosed |
| License | Apache License 2.0 |
| Modalities | Text input Image input Video input Text output |
| Tags | Agentic Small LM Tool use Vision |
| End of life | Not disclosed |
| Weights | Download |
| Last verified | Not disclosed |