InternVL3-14B-Instruct: benchmark results
Provider: Shanghai AI Lab. Access: Open.
Aggregate Rating 1542 ± 35, rank #1813 of 10916 rated models, from 16 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| MCSBench - L2 | 57.79 | Accuracy (%) | 71.1 |
| MCSBench - Overall | 51.59 | Accuracy (%) | 67.1 |
| MCSBench - L1 | 62.07 | Accuracy (%) | 60.5 |
| MCSBench - L3 | 23.66 | Accuracy (%) | 55.3 |
| AICA-Bench | 62.41 | Overall average on a percent scale: unweighted mean of the e | 54.5 |
| AICA-Bench - Emotion-Guided Generation | 66.5 | Emotion-guided content generation score on a percent scale: | 54.5 |
| SciVQR - Geography | 60.7 | Geography accuracy (%) with zero-shot chain-of-thought promp | 50 |
| SciVQR | 49.9 | Overall accuracy (%) with zero-shot chain-of-thought prompti | 42.9 |
| SciVQR - Astronomy | 55.5 | Astronomy accuracy (%) with zero-shot chain-of-thought promp | 42.9 |
| SciVQR - Biology | 57.2 | Biology accuracy (%) with zero-shot chain-of-thought prompti | 42.9 |
| SciVQR - Chemistry | 49.6 | Chemistry accuracy (%) with zero-shot chain-of-thought promp | 42.9 |
| SciVQR - Mathematics | 36.9 | Mathematics accuracy (%) with zero-shot chain-of-thought pro | 42.9 |
Interactive version: theaggregate.ai/model?slug=internvl3-14b-instruct · How It Works · Data refreshed daily, snapshot 2026-11-08.