Model
GPT-oss 120B
OpenAI · open weights
Knowledge & Reasoning
Knowledge & Reasoning| Benchmark | Score | Metric | Harness | Source | Confidence | Link |
|---|---|---|---|---|---|---|
| CritPt | 1.1 | accuracy | N/A | Independent | low | epoch.ai |
| GPQA Diamond | 80.9 | accuracy | N/A | Self-reported | low | artificialanalysis.ai |
| GPQA Diamond | 75.8 | accuracy | N/A | Independent | high | epoch.ai |
| LMArena | 1355.0 | elo | N/A | Self-reported | low | lmarena.ai |
| MMLU-Pro | 90.0 | accuracy | N/A | Self-reported | low | artificialanalysis.ai |
| PARCEL | 90.0 | accuracy | N/A | Independent | medium | arxiv.org |
Math
Math| Benchmark | Score | Metric | Harness | Source | Confidence | Link |
|---|---|---|---|---|---|---|
| AIME 2025 | 97.9 | accuracy | N/A | Self-reported | low | kaggle.com |
| SELF-POT | 0.5 | success_rate | N/A | Independent | medium | arxiv.org |
Coding
Coding| Benchmark | Score | Metric | Harness | Source | Confidence | Link |
|---|---|---|---|---|---|---|
| Aider Polyglot | 41.8 | percent_correct | N/A | Independent | medium | epoch.ai |
| AptMQL-Bench | 62.5 | soft-ex | N/A | Independent | medium | arxiv.org |
| LiveCodeBench | 60.0 | pass@1 | N/A | Self-reported | low | livecodebench.github.io |
| SciCode | 36.0 | score | N/A | Independent | medium | artificialanalysis.ai |
| SWE-bench Verified | 62.4 | resolved_rate | N/A | Self-reported | low | swebench.com |
| Terminal-Bench 2.0 | 18.7 | accuracy | N/A | Self-reported | low | tbench.ai |
Agentic / Tool Use
Agentic / Tool Use| Benchmark | Score | Metric | Harness | Source | Confidence | Link |
|---|---|---|---|---|---|---|
| DelegationBench | 54.5 | accuracy | N/A | Independent | medium | arxiv.org |
| DyadMem | 61.2 | accuracy | N/A | Independent | medium | arxiv.org |
| GeoNatureAgent (GNA) | 34.1 | accuracy | N/A | Independent | medium | arxiv.org |
| ReFract | 0.3 | success_rate | function calling | Independent | medium | arxiv.org |
| Terminal-Bench | 18.7 | accuracy | Terminus 2 | Independent | medium | tbench.ai |
| Terminal-Bench | 14.2 | accuracy | Mini-SWE-Agent | Independent | medium | tbench.ai |
| XiangqiBench | 0.8 | pass@3 | N/A | Independent | medium | arxiv.org |
Safety & Alignment
Safety & Alignment| Benchmark | Score | Metric | Harness | Source | Confidence | Link |
|---|---|---|---|---|---|---|
| RobustReview | 0.5 | joint_stability-discrimination_profile_and_human_alignment_scores | N/A | Independent | medium | arxiv.org |
Abstract Reasoning
Abstract Reasoning| Benchmark | Score | Metric | Harness | Source | Confidence | Link |
|---|---|---|---|---|---|---|
| Actionable Indeterminacy Benchmark | 83.5 | accuracy | N/A | Independent | medium | arxiv.org |
| ArgGYM | 78.1 | accuracy | N/A | Independent | medium | arxiv.org |
| auto-nesy-bench | 98.4 | f1 | zs-cot | Independent | medium | arxiv.org |
Reproduced by the community
Loading community runs…