Back to BenchmarksVery Hard
PIQA
Physical intuition question answering
Leaderboard
(16 models)| Rank | Model | Score | Stderr |
|---|---|---|---|
| 1 | gpt-4o-mini-2024-07-18 | 88.70 | — |
| 2 | Phi-3-mini-4k-instruct | 88.60 | — |
| 3 | Gemini 1.5 Flash | 87.50 | — |
| 4 | Llama 3.1 405B | 85.90 | — |
| 5 | falcon-180B | 84.90 | — |
| 6 | DeepSeek V3 | 84.70 | — |
| 7 | Gemma 3 27B | 83.70 | — |
| 8 | Mixtral-8x7B-v0.1 | 83.60 | — |
| 9 | Mistral Large | 83.50 | — |
| 10 | Mistral-7B-v0.1 | 83.00 | — |
| 11 | Llama-2-70b-hf | 82.80 | — |
| 12 | Qwen 2.5 72B | 82.60 | — |
| 13 | Llama-2-7b | 81.90 | — |
| 14 | gemma-7b | 81.20 | — |
| 15 | Qwen 3 235B | 79.90 | — |
| 16 | GPT-OSS 120B | 76.70 | — |
Data source: Epoch AI, “Data on AI Benchmarking”. Published at epoch.ai
Licensed under CC-BY 4.0