GPT-OSS-Swallow-120B-SFT-v0.1: benchmark results
Provider: OpenAI. Access: API.
Aggregate Rating 1619 ± 33, rank #845 of 10916 rated models, from 35 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Swallow - English MT-Bench - Math | 99.1 | Judge Score (normalized, %) | 86.6 |
| Swallow - Post-trained Japanese - WMT20 En-Ja | 26.4 | BLEU | 86.6 |
| Swallow - Japanese MT-Bench - Math | 98.7 | Judge Score (normalized, %) | 85.1 |
| Swallow - Japanese MT-Bench - Coding | 80.9 | Judge Score (normalized, %) | 84.3 |
| Swallow - Post-trained Japanese - M-IFEval-Ja | 74.8 | Instruction-Level Strict Accuracy (%) | 82.1 |
| Swallow - Japanese MT-Bench - Writing | 68.7 | Judge Score (normalized, %) | 80.6 |
| Swallow - English MT-Bench - Extraction | 81.4 | Judge Score (normalized, %) | 79.1 |
| Swallow - Post-trained English - IFBench | 56.4 | Instruction-Level Strict Accuracy (%) | 79.1 |
| Swallow - Japanese MT-Bench - Average | 75.8 | Judge Score (normalized, %) | 78.4 |
| Swallow - Post-trained Japanese - JHumanEval | 92.7 | Pass@1 (%) | 78.4 |
| Swallow - English MT-Bench - Roleplay | 76.9 | Judge Score (normalized, %) | 77.6 |
| Swallow - English MT-Bench - Writing | 75.6 | Judge Score (normalized, %) | 77.6 |
Interactive version: theaggregate.ai/model?slug=gpt-oss-swallow-120b-sft-v0-1 · How It Works · Data refreshed daily, snapshot 2026-11-08.