Qwen3-VL-Plus-2025-09-23
Alibaba · 2025-09-23
Qwen3-VL-Plus-2025-09-23 is Alibaba Cloud Model Studio's closed-weight Qwen3-VL-Plus snapshot released September 23, 2025. It accepts text, image, and video and returns text within a 262,144-token context window; its hybrid thinking mode can be enabled or disabled per invocation. Alibaba describes upgrades in visual coding, spatial perception, multimodal reasoning, visual perception and recognition, and ultra-long-video understanding.
Benchmark scores
| Benchmark | Score |
|---|---|
| BabyVision | 19.2 |
| Capability | 142.3 |
| MedCode | 31.7 |
| MedScribe | 77.1 |
| SAGE | 42.1 |
| SAGE - Automata | 45.6 |
| SAGE - Calculus | 15.4 |
| SAGE - Linear Algebra | 53.7 |
| SAGE - Partial Differential Equations | 27.0 |