Qwen 3 VL 30B: benchmark results
Provider: Alibaba.
Aggregate Rating 1666 ± 27, rank #520 of 10916 rated models, from 68 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| EcoG-Bench (Frames + ASR) | 11.4 | Eco-accuracy (%): share of referents whose option (what), 2D | 100 |
| EcoG-Bench (Frames + ASR) - Sequence Success | 6.7 | Sequence success (%): share of clips in which every referent | 100 |
| MMOOC - Image-Question Mismatch (Multiple Choice) | 95 | Shifted in-context score (%; mean of accuracy on the answera | 100 |
| MMOOC - Visual False Premises (Multiple Choice) | 36.25 | Out-of-context score (%; mean of the refusal rate, the share | 100 |
| GeoMMBench - LiDAR | 80.6 | Micro-averaged accuracy (%) on the LiDAR questions of the Ge | 98.5 |
| GeoMMBench - Time Series | 63 | Micro-averaged accuracy (%) on the time-series questions of | 98.5 |
| GeoMMBench - GNSS | 76.5 | Micro-averaged accuracy (%) on the global navigation satelli | 97.1 |
| MMOOC - Image-Question Mismatch (Yes/No) | 78.5 | Shifted in-context score (%; mean of accuracy on the answera | 94.1 |
| MMOOC - Misleading Premise (Yes/No) | 88.5 | Shifted in-context score (%; mean of accuracy on the answera | 94.1 |
| GeoMMBench | 66.7 | Micro-averaged accuracy (%) on all questions of the GeoMMBen | 91.2 |
| GeoMMBench - Applications | 67.8 | Micro-averaged accuracy (%) on the remote sensing applicatio | 91.2 |
| GeoMMBench - Data Quality | 65.4 | Micro-averaged accuracy (%) on the data quality questions (i | 91.2 |
Interactive version: theaggregate.ai/model?slug=qwen-3-vl-30b · How It Works · Data refreshed daily, snapshot 2026-11-08.