Open LLM Leaderboard v1 - GSM8K: leaderboard
Original Hugging Face Open LLM Leaderboard grade-school arithmetic word problems, with five demonstrations and the harness answer-accuracy metric. Percentage, higher is better. Frozen v1 archive; genuine zero scores are retained and public questions may overlap model training.
Metric: Accuracy (%) (5-shot). Source: huggingface.co. Saturation forecast: Estimated already saturated. 7252 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Llama 3 70B Japanese Suzume Vector V0.1 | 87.41 | #774 |
| 2 | Llama3-TenyxChat-70B | 86.28 | #739 |
| 3 | autotrain-llama3-70B-orpo-v1 | 85.67 | #844 |
| 4 | Llama 3 70B Instruct | 85.44 | #643 |
| 5 | Qwen 2 72B | 85.29 | #517 |
| 6 | Llama 3 70B Fireplace | 83.62 | #847 |
| 7 | Mixtral 8x22B Instruct | 82.03 | #741 |
| 8 | Qwen 1.5 110B | 81.05 | #735 |
| 9 | Le_Triomphant-ECE-TW3 | 79.83 | #564 |
| 10 | Smaug-72B-v0.1 | 78.7 | #613 |
| 11 | Llama 3 70B | 76.88 | #635 |
| 12 | Rhea-72B-v0.5 | 76.12 | #555 |
| 13 | free-evo-qwen72B-v0.8-re | 75.89 | #442 |
| 14 | YiSM-34B-0rn | 75.82 | #715 |
| 15 | SeaLLM-7B-v2.5 | 74.37 | #1067 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #6 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=open-llm-leaderboard-v1-gsm8k · How It Works · Data refreshed daily, snapshot 2026-11-07.