Solar Pro 3: benchmark results
Upstage's enterprise reasoning MoE (102B total/12B active), Korean/English bilingual (January 2026). Provider: Upstage. Released 2026-01-26. Access: API.
Aggregate Rating 1566 ± 19, rank #557 of 1656 rated models, from 105 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AA IFBench | 71.2 | Accuracy (%) | 86.9 |
| AA TAU-2 Bench | 86.26 | Accuracy (%) | 78.9 |
| Tau2-Bench Telecom | 86.26 | Success Rate (%) | 78.7 |
| KoNeoBench - Definition with Examples | 7.79 | Definition Judge Score (0-10) | 75 |
| KoNeoBench - Definition without Examples | 6.68 | Definition Judge Score (0-10) | 75 |
| KoNeoBench - Source Words - Coverage | 32.48 | Coverage (%) | 75 |
| KoNeoBench - Source Words - F1 | 32.9 | F1 (%) | 75 |
| KoNeoBench - Source Words - Precision | 33.86 | Precision (%) | 75 |
| KoNeoBench - Specialist Domain - Open without Examples | 49.38 | Accuracy (%) | 62.5 |
| UGI - Natural Intelligence | 23.04 | NatInt Score | 53.5 |
| AA GPQA Diamond | 72.42 | Accuracy (%) | 51.8 |
| KoNeoBench - Semantic Category - Multiple Choice | 37.14 | Accuracy (%) | 50 |
Interactive version: theaggregate.ai/model?slug=solar-pro-3 · How It Works · Data refreshed daily, snapshot 2026-10-20.