Arcadia CommonsenseQA: leaderboard
Arcadia Impact inspect_ai evaluation of CommonsenseQA, testing commonsense multiple-choice reasoning over everyday concepts and relations.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 11 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Grok 2 (1212) | 86.24 | #496 |
| 2 | GPT-4o (2024-08-06) | 85.91 | #336 |
| 3 | Qwen 2.5 VL 72B Instruct | 85.59 | #377 |
| 4 | Claude 3.7 Sonnet (20250219) | 84.03 | #201 |
| 5 | DeepSeek V3 Chat | 84.03 | #447 |
| 6 | Llama 3.3 70B Instruct | 83.37 | #534 |
| 7 | Llama 3.2 90B Vision Instruct | 82.8 | #622 |
| 8 | Gemini 2.0 Flash (001) | 82.47 | #360 |
| 9 | GPT-4o Mini (2024-07-18) | 81.98 | #558 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=arcadia-commonsenseqa · How It Works · Data refreshed daily, snapshot 2026-10-20.