HELM Classic - MATH Chain-of-Thought: leaderboard
The "MATH Chain-of-Thought" subtask of HELM Classic. A scenario from Stanford HELM Classic, the holistic evaluation suite scoring language models across accuracy, robustness and calibration on a broad set of tasks.
Metric: Equivalent (%). Source: crfm.stanford.edu. Saturation forecast: Estimated already saturated. 69 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-3.5 Turbo (0613) | 71.86 | #1118 |
| 2 | text-davinci-003 | 44.92 | #1146 |
| 3 | text-davinci-002 | 38.06 | #1225 |
| 4 | Palmyra X (43B) | 38.02 | #1317 |
| 5 | Llama 2 70B | 33 | #1201 |
| 6 | Mistral-7B-v0.1 | 29.34 | #1326 |
| 7 | LLaMA-65B | 25.25 | #1303 |
| 8 | LLaMA-30B | 22.29 | #1423 |
| 9 | falcon-40B Instruct | 13.65 | #1458 |
| 10 | falcon-40B | 13.01 | #1442 |
| 11 | mpt-30B Instruct | 12.54 | #1470 |
| 12 | mpt-30B | 12.41 | #1476 |
| 13 | vicuna-13B-v1.3 | 12.25 | #1447 |
| 14 | Llama 2 13B | 12.18 | #1436 |
| 15 | LLaMA-13B | 10.98 | #1516 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=helm-classic-math-chain-of-thought · How It Works · Data refreshed daily, snapshot 2026-10-20.