MMLU-by-task - Astronomy — leaderboard

Metric: Accuracy (%). Source: huggingface.co. 1257 models tracked.

Top models

#ModelScore
1Llama 2 70B Base80.92
2Llama 2 70B Chat79.61
3StableBeluga278.95
4falcon-180B78.95
5Llama 2 70B Chat (HF)73.03
6Llama 2 70B Chat GPTQ73.03
7LLaMA-65B71.71
8Mistral-7B-v0.165.79
9vicuna-33B-v1.365.13
10LLaMA-30B61.18
11internlm-20B Chat57.89
12OpenHermes-13B57.24
13trurl-2-13B-academic56.58
14vicuna-13B-v1.555.92
15falcon-40B Instruct55.26

Interactive version: theaggregate.ai/benchmark?slug=mmlu-by-task-astronomy · How the rankings work · Data refreshed daily, snapshot 2026-07-22.