Open CoT - LSAT Analytical Reasoning — leaderboard

Metric: CoT Gain (%). Source: huggingface.co. 132 models tracked.

Top models

#ModelScore
1DeepSeek R1 Distill Llama 70B22.61
2DeepSeek R1 Distill Qwen 14B16.96
3DeepSeek R1 Distill Qwen 32B16.52
4Llama-3.1-SuperNova-Lite13.91
5Llama 3.1 8B Instruct11.74
6Qwen 2.5 14B Instruct11.74
7merlinite-7B10.43
8Qwen 2.5 7B Instruct9.13
9Llama 3.1 Nemotron 70B Instruct8.7
10Mixtral 8x7B Instruct (v0.1)8.7
11SuperNova-Medius8.7
12Llama 3 8B Instruct8.26
13DeepSeek R1 Distill Llama 8B8.26
14Llama 3 70B8.26
15internlm2-chat-20B8.26

Interactive version: theaggregate.ai/benchmark?slug=open-cot-lsat-analytical-reasoning · How the rankings work · Data refreshed daily, snapshot 2026-07-22.