GPT-OSS-120B (Medium): benchmark results
GPT-OSS-120B evaluated at the medium reasoning-effort setting. Provider: OpenAI. Released 2025-08-05. Access: Open.
Aggregate Rating 1617 ± 9, rank #655 of 2168 rated models, from 169 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Multi-LCB - Pass@5 T0.6 - Go | 85.1 | Pass@5 (%) | 100 |
| Multi-LCB - Pass@5 T0.6 - JavaScript | 84.1 | Pass@5 (%) | 100 |
| Multi-LCB - Pass@5 T0.6 - Kotlin | 83.7 | Pass@5 (%) | 100 |
| Multi-LCB - Pass@5 T0.6 - Overall | 83.6 | Pass@5 (%) | 100 |
| Multi-LCB - Pass@5 T0.6 - PHP | 82.1 | Pass@5 (%) | 100 |
| Multi-LCB - Pass@5 T0.6 - Ruby | 83.3 | Pass@5 (%) | 100 |
| Multi-LCB - Pass@5 T0.6 - Rust | 84.1 | Pass@5 (%) | 100 |
| Multi-LCB - Pass@5 T0.6 - Scala | 80.2 | Pass@5 (%) | 100 |
| Multi-LCB - Pass@5 T0.6 - TypeScript | 85.2 | Pass@5 (%) | 100 |
| Rule2DRC | 32.5 | Success rate (%; one sampled DRC script per rule with the 60 | 100 |
| Multi-LCB - Pass@5 T0.6 - C# | 80.7 | Pass@5 (%) | 95.8 |
| Multi-LCB - Pass@5 T0.6 - C++ | 84.9 | Pass@5 (%) | 95.8 |
Interactive version: theaggregate.ai/model?slug=gpt-oss-120b-medium · How It Works · Data refreshed daily, snapshot 2026-10-20.