Intern-S1-Pro: benchmark results
Provider: Shanghai AI Lab. Access: Open.
Aggregate Rating 1654 ± 23, rank #598 of 10916 rated models, from 30 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| SciHazard (Science-Specific Models) | 0.88 | DeHarm-Score (0-5, lower is safer; the mean over 2,400 hazar | 100 |
| LithoBench | 93.38 | MCQ$_{\mathrm{all}}$ (self-reported) | 80 |
| LithoBench - Identification and Description (Open-Ended) | 3.23 | OPEN_J score (0-10) = 0.4 x ROUGE-L x 10 + 0.2 x BLEU-4 x 10 | 80 |
| LithoBench - Practical Application (Open-Ended) | 3.91 | OPEN_J score (0-10) = 0.4 x ROUGE-L x 10 + 0.2 x BLEU-4 x 10 | 80 |
| LiveSecBench | 63.63 | Overall Score (%) | 73.8 |
| LithoBench - Comprehensive Reasoning (Open-Ended) | 4.42 | OPEN_J score (0-10) = 0.4 x ROUGE-L x 10 + 0.2 x BLEU-4 x 10 | 70 |
| LithoBench - Open-Ended | 3.88 | OPEN_J score (0-10) = 0.4 x ROUGE-L x 10 + 0.2 x BLEU-4 x 10 | 70 |
| OmniMatBench - Calculation | 23.2 | Calculation strict exact-slot accuracy (%): share of structu | 66.7 |
| LithoBench - Identification and Description (MCQ) | 70.38 | Accuracy (%) on the 800 multiple-choice identification-and-d | 63.6 |
| LithoBench - Comparative Analysis (Open-Ended) | 3.73 | OPEN_J score (0-10) = 0.4 x ROUGE-L x 10 + 0.2 x BLEU-4 x 10 | 60 |
| LithoBench - Mechanism Explanation (Open-Ended) | 4.1 | OPEN_J score (0-10) = 0.4 x ROUGE-L x 10 + 0.2 x BLEU-4 x 10 | 60 |
| UAV-DualCog - Self-Relative Position | 29.4 | Answer accuracy (%; 1,024 questions asking where the UAV is | 51.4 |
Interactive version: theaggregate.ai/model?slug=intern-s1-pro · How It Works · Data refreshed daily, snapshot 2026-11-08.