AI models
gpt-oss-120b vs Qwen3.8 27B
Same criteria, side by side. Where the gap is small, it may be measurement noise.
Also searched as
gpt-oss vs Qwen3.8
Direct answer: GPQA Diamond (OpenRouter): gpt-oss-120b 77.4%, Qwen3.8 27B 87.2% (OpenRouter Benchmarks). τ-bench Verified (airline): gpt-oss-120b 67.3%, Qwen3.8 27B 78% (OpenRouter Benchmarks).
| Metric | gpt-oss-120b | Qwen3.8 27B | Source |
|---|---|---|---|
| Weights ?what it is | open | open | no data |
| Input US$/1M ? | US$ 0.15 | US$ 0.15 | no data |
| Output US$/1M ? | US$ 0.60 | US$ 1.88 | no data |
| Context ? | 131,072 | 1,048,575 | no data |
| GPQA Diamond ? | 75.8% | no data | Epoch AI |
| OTIS Mock AIME ? | 88.9% | no data | Epoch AI |
| GPQA Diamond (OpenRouter) ? | 77.4% | 87.2% | OpenRouter Benchmarks |
| τ-bench Verified (airline) ? | 67.3% | 78% | OpenRouter Benchmarks |
Compare with
Sources
- Epoch AI, CC BY 4.0, latest version we checked.
- OpenRouter Benchmarks, CC BY 4.0, latest version we checked.
- OpenRouter, dados públicos da API; exports de benchmarks e rankings em CC BY 4.0, latest version we checked.
Dates come from the sources: OpenRouter listing or Hugging Face repo creation, which may differ from the official launch. Family and version were grouped by rule from the name.