Solar Pro 2 (Reasoning) — benchmark results
Provider: Upstage. Released 2025-05-20. Access: Open.
Unified ELO 1550 ± 37, rank #648 of 1841 rated models, from 54 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AA MATH-500 | 96.67 | Accuracy (%) | 85.4 |
| AA MMLU-Pro | 80.52 | Accuracy (%) | 69.2 |
| AA LiveCodeBench | 61.59 | Pass@1 (%) | 67.3 |
| AA Omniscience - Software Engineering (SWE) - HTML | 36 | Accuracy (%) | 62.3 |
| AA Omniscience - Health | 21.3 | Accuracy (%) | 58.7 |
| AA Omniscience - Science, Engineering & Mathematics | 28.1 | Accuracy (%) | 58.1 |
| AA AIME 2025 | 61.33 | Accuracy (%) | 57.4 |
| AA Global-MMLU-Lite - Indonesian | 86.25 | Accuracy (%) | 55.6 |
| AA Omniscience - Software Engineering (SWE) - Swift | 40 | Accuracy (%) | 55.4 |
| AA Global-MMLU-Lite - Chinese | 85.83 | Accuracy (%) | 52.7 |
| AA Global-MMLU-Lite - Japanese | 85.5 | Accuracy (%) | 52.6 |
| AA Omniscience - Law | 11.3 | Accuracy (%) | 51.3 |
Interactive version: theaggregate.ai/model?slug=solar-pro-2-reasoning · How It Works · Data refreshed daily, snapshot 2026-07-25.