PIQA: leaderboard
Physical Intuition QA: 20K multiple-choice questions testing physical commonsense reasoning - understanding how everyday objects and actions work in the physical world.
Metric: Accuracy (%). Source: epoch.ai. Saturation forecast: Estimated already saturated. 60 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Median Human | 94.9 | |
| 2 | GPT-4o Mini (2024-07-18) | 88.7 | #558 |
| 3 | Phi-3.5-MoE-instruct | 88.6 | #700 |
| 4 | Gemini 1.5 Flash (002) | 87.5 | #583 |
| 5 | Llama 3.1 405B | 85.9 | #466 |
| 6 | DeepSeek V2 | 83.9 | #549 |
| 7 | Gemma 2 9B | 83.7 | #958 |
| 8 | Mixtral 8x7B (v0.1) | 83.6 | #1083 |
| 9 | Mistral-Nemo-Base-2407 | 83.5 | #1228 |
| 10 | StableBeluga2 | 83.3 | #1078 |
| 11 | Mistral-7B-v0.1 | 83 | #1326 |
| 12 | falcon-40B | 83 | #1442 |
| 13 | Llama 2 70B Base | 82.8 | #1025 |
| 14 | LLaMA-65B | 82.8 | #1303 |
| 15 | Qwen 2.5 72B | 82.6 | #499 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=piqa · How It Works · Data refreshed daily, snapshot 2026-10-20.