InternVL2-76B: benchmark results
Provider: Shanghai AI Lab. Released 2024-07-15. Access: Open.
Aggregate Rating 1559 ± 20, rank #1559 of 10916 rated models, from 634 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| MEGA-Bench Task - Signboard Identification | 93.3 | Task Score (%) | 100 |
| MEGA-Bench Task - Symbolic Graphics Programs Scalable Vector Graphics | 27.8 | Task Score (%) | 100 |
| MEGA-Bench Task - Visual Prediction Rater Openable Part Segmentation | 42.9 | Task Score (%) | 100 |
| OpenVLM Video - MVBench - Fine-grained Pose | 86 | Accuracy (%) | 100 |
| MEGA-Bench Task - Autorater Mask | 78.6 | Task Score (%) | 97.7 |
| MEGA-Bench Task - Planning Screenshot Tyreworld | 100 | Task Score (%) | 97.7 |
| OpenVLM Video - MVBench - Fine-grained Action | 55 | Accuracy (%) | 97.2 |
| MEGA-Bench Task - Bongard Problem | 36.8 | Task Score (%) | 96.2 |
| MEGA-Bench Task - Monthly Weather Days Count | 45.2 | Task Score (%) | 94.2 |
| MEGA-Bench Task - Photo Sharing Image Retrieval | 92.9 | Task Score (%) | 94.2 |
| MEGA-Bench Task - Super CLEVR | 85.7 | Task Score (%) | 94.2 |
| MEGA-Bench Task - YouTube Video Info Parsing | 82.1 | Task Score (%) | 94.2 |
Interactive version: theaggregate.ai/model?slug=internvl2-76b · How It Works · Data refreshed daily, snapshot 2026-11-08.