Model
GPT-3.5 Turbo
OpenAI · closed weights
Knowledge & Reasoning
Knowledge & Reasoning| Benchmark | Score | Metric | Harness | Source | Confidence | Link |
|---|---|---|---|---|---|---|
| ARC-Challenge | 85.2 | accuracy | N/A | Independent | medium | arxiv.org |
| EconLogicQA | 37.7 | accuracy | N/A | Independent | medium | arxiv.org |
| HellaSwag | 85.5 | accuracy | N/A | Independent | medium | arxiv.org |
| LegalBench | 60.9 | balanced_accuracy | N/A | Independent | medium | paperswithcode.com |
| MedConceptsQA | 41.5 | accuracy | N/A | Independent | medium | arxiv.org |
| MMLU | 70.0 | average_(%) | N/A | Independent | medium | arxiv.org |
| MMLU | 68.5 | exact_match | N/A | Independent | medium | arxiv.org |
| OCW | 82.5 | wasserstein_distance_(wd) | N/A | Independent | medium | arxiv.org |
| OCW | 82.3 | wasserstein_distance_(wd) | N/A | Independent | medium | arxiv.org |
| OCW | 81.2 | wasserstein_distance_(wd) | N/A | Independent | medium | arxiv.org |
| OCW | 80.9 | wasserstein_distance_(wd) | N/A | Independent | medium | arxiv.org |
| OCW | 80.6 | wasserstein_distance_(wd) | N/A | Independent | medium | arxiv.org |
| WinoGrande | 81.6 | accuracy | N/A | Independent | medium | arxiv.org |
Math
MathCoding
Coding| Benchmark | Score | Metric | Harness | Source | Confidence | Link |
|---|---|---|---|---|---|---|
| MBPP | 70.8 | accuracy | N/A | Independent | medium | arxiv.org |
| MBPP | 67.6 | accuracy | N/A | Independent | medium | arxiv.org |
| MBPP | 52.2 | accuracy | N/A | Independent | medium | arxiv.org |
| MBPP | 45.4 | accuracy | N/A | Independent | medium | arxiv.org |
| MBPP | 39.8 | accuracy | N/A | Independent | medium | arxiv.org |
| MMSQL | 64.1 | tdex | N/A | Independent | medium | arxiv.org |
| PECC | 23.8 | pass@3 | N/A | Independent | medium | arxiv.org |
| Turbulence | 61.7 | corrsc | N/A | Independent | medium | arxiv.org |
Safety & Alignment
Safety & AlignmentMultilingual
MultilingualReproduced by the community
Loading community runs…