Qwen 2 VL 2B Instruct: benchmark results
Provider: Alibaba. Released 2024-08-29. Access: Open.
Aggregate Rating 1423 ± 19, rank #5028 of 10916 rated models, from 86 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| MERA Multi - UniScienceVQA | 11.2 | Score (%) | 69.5 |
| MERA Multi - ruMathVQA | 4.5 | Score (%) | 66.1 |
| MERA Multi - ruCLEVR | 22 | Score (%) | 59.3 |
| MERA Multi - ruNaturalScienceVQA | 18.3 | Score (%) | 57.6 |
| MERA Multi - RealVQA | 16.8 | Score (%) | 52.5 |
| MERA Multi - SchoolScienceVQA | 20.7 | Score (%) | 52.5 |
| FewMMBench (8-Shot, Similar + CoT) | 64 | Mean accuracy (%) on the FewMMBench query sets (about 250 mu | 52 |
| MERA Multi - ruTiE-Image | 29.8 | Score (%) | 50 |
| KOFFVQA - Document Understanding | 54 | Score (%) | 48.1 |
| FewMMBench (4-Shot, Similar + CoT) | 64 | Mean accuracy (%) on the FewMMBench query sets (about 250 mu | 48 |
| MERA Multi - Image | 16.5 | Score (%) | 45.8 |
| EuraGovExam - Psychology | 27.73 | Accuracy (%) on the psychology questions across the five reg | 44.4 |
Interactive version: theaggregate.ai/model?slug=qwen-2-vl-2b-instruct · How It Works · Data refreshed daily, snapshot 2026-11-08.