HELM Lite - MATH Level 1 - Prealgebra: leaderboard
HELM Lite v1.13.0 MATH chain-of-thought run on Level 1 (easiest) Prealgebra problems only: 86 test problems, up to 8-shot, scored by equivalence of the final answer to the reference. Percent, 0-100, higher is better; a small sample, not comparable with full-MATH Prealgebra scores.
Metric: Equivalent (CoT) (%). Source: crfm.stanford.edu. Saturation forecast: Estimated already saturated. 91 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude 3.5 Haiku (20241022) | 98.84 | #590 |
| 2 | Claude 3.5 Sonnet (20241022) | 97.67 | #302 |
| 3 | GPT-4o (2024-05-13) | 97.67 | #337 |
| 4 | Qwen 2.5 72B Instruct | 96.51 | #469 |
| 5 | Gemini 2.0 Flash (Preview) | 96.51 | #376 |
| 6 | Gemini 1.5 Flash (002) | 95.35 | #587 |
| 7 | Claude 3.5 Sonnet (20240620) | 95.35 | #308 |
| 8 | GPT-4o Mini (2024-07-18) | 95.35 | #568 |
| 9 | GPT-4 Turbo | 95.35 | #417 |
| 10 | Gemini 1.5 Pro (001) | 95.35 | #465 |
| 11 | Llama 3.3 70B Instruct | 94.19 | #541 |
| 12 | Gemini 1.5 Pro (002) | 94.19 | #368 |
| 13 | DeepSeek V3 | 94.19 | #329 |
| 14 | GPT-4o (2024-08-06) | 93.02 | #343 |
| 15 | Gemma 2 27B (IT) | 93.02 | #652 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=helm-lite-math-level-1-prealgebra · How It Works · Data refreshed daily, snapshot 2026-11-05.