HELM MMLU - Astronomy: leaderboard
The "Astronomy" subtask of HELM MMLU. HELM Classic MMLU group leaderboard from the public HELM release artifacts.
Metric: EM. Source: crfm.stanford.edu. Saturation forecast: Estimated already saturated. 79 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude 3.5 Sonnet (20241022) | 97.37 | #291 |
| 2 | Claude 3 Opus (20240229) | 96.71 | #413 |
| 3 | Claude 3.5 Sonnet (20240620) | 96.05 | #298 |
| 4 | GPT-4o (2024-08-06) | 94.74 | #335 |
| 5 | GPT-4o (2024-05-13) | 94.08 | #326 |
| 6 | GPT-4 Turbo | 94.08 | #406 |
| 7 | GPT-4 Turbo (Preview) | 94.08 | #399 |
| 8 | Qwen 2.5 72B Instruct | 93.42 | #458 |
| 9 | Gemini 1.5 Pro (002) | 93.42 | #350 |
| 10 | Qwen 2 72B Instruct | 93.42 | #579 |
| 11 | GPT-4 (0613) | 93.42 | #494 |
| 12 | Gemini 2.0 Flash (Preview) | 92.76 | #366 |
| 13 | DeepSeek V3 | 92.11 | #323 |
| 14 | Llama 3.1 405B Instruct | 92.11 | #461 |
| 15 | Mistral Large 2 (Jul) | 92.11 | #485 |
No result here: #4 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=helm-mmlu-astronomy · How It Works · Data refreshed daily, snapshot 2026-10-23.