InternVL3-1B-Instruct: benchmark results
Provider: Shanghai AI Lab. Access: Open.
Aggregate Rating 1227 ± 51, rank #9436 of 10916 rated models, from 23 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| MERA Multi - WEIRD | 2.8 | Score (%) | 19.5 |
| MERA Multi - ruCommonVQA | 3.8 | Score (%) | 18.6 |
| MERA Multi - ruCLEVR | 0.7 | Score (%) | 14.4 |
| MERA Multi - ruMathVQA | 0.5 | Score (%) | 14.4 |
| MERA Multi - RealVQA | 0.8 | Score (%) | 12.7 |
| MERA Multi - ruTiE-Image | 0.7 | Score (%) | 12.1 |
| FewMMBench (Zero-Shot) | 39.3 | Mean accuracy (%) on the FewMMBench query sets (about 250 mu | 12 |
| MERA Multi - ruHHH-Image | 0 | Score (%) | 11 |
| MERA Multi - Image | 0.9 | Score (%) | 8.5 |
| FewMMBench (4-Shot, Random) | 45.8 | Mean accuracy (%) on the FewMMBench query sets (about 250 mu | 8 |
| FewMMBench (4-Shot, Similar + CoT) | 34.3 | Mean accuracy (%) on the FewMMBench query sets (about 250 mu | 8 |
| FewMMBench (8-Shot, Random) | 45.7 | Mean accuracy (%) on the FewMMBench query sets (about 250 mu | 8 |
Interactive version: theaggregate.ai/model?slug=internvl3-1b-instruct · How It Works · Data refreshed daily, snapshot 2026-11-08.