INTELLECT-3 (Based on GLM-4.5-Air): benchmark results
Provider: PrimeIntellect.
Aggregate Rating 1626 ± 24, rank #788 of 10916 rated models, from 27 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AA LiveCodeBench | 77.67 | Pass@1 (%) | 89.3 |
| AA AIME 2025 | 88 | Accuracy (%) | 86 |
| AA MMLU-Pro | 82.24 | Accuracy (%) | 79.3 |
| AA GPQA Diamond | 76.06 | Accuracy (%) | 58.9 |
| AA Humanity's Last Exam | 13.07 | Accuracy (%) | 56.5 |
| AA Global-MMLU-Lite - Portuguese | 86.67 | Accuracy (%) | 49.1 |
| AA Global-MMLU-Lite - Bengali | 80.58 | Accuracy (%) | 47.3 |
| AA Global-MMLU-Lite - French | 86.25 | Accuracy (%) | 46.4 |
| AA Global-MMLU-Lite - Italian | 87.25 | Accuracy (%) | 46.4 |
| AA CritPt | 0.29 | Accuracy (%) | 45.9 |
| AA Global-MMLU-Lite - Burmese | 71.75 | Accuracy (%) | 45.5 |
| AA Global-MMLU-Lite - Arabic | 83.08 | Accuracy (%) | 43.6 |
Interactive version: theaggregate.ai/model?slug=intellect-3-based-on-glm-4-5-air · How It Works · Data refreshed daily, snapshot 2026-11-08.