General · %
Open LLM Leaderboard v2
Hugging Face's leaderboard of open-weight models, scored on IFEval, BBH, MATH Lvl 5, GPQA, MuSR and MMLU-Pro with normalised scores. The leaderboard was archived in 2025 and no longer accepts submissions.
Archived Contamination risk: High 311 results · 294 models Updated Oct 22, 2026 Newest result: 2025-03-13 · stale Official leaderboard ↗
What it measures
Average of six normalised academic benchmarks, run by Hugging Face with the same harness.
Limitations
Archived: no models after early 2025. Only open-weight models. We keep only models published by their original developers (no merges or community fine-tunes).
Contamination risk: High
Fixed public test sets; the leaderboard itself flagged contaminated submissions.
Leaderboard
Top 100 of 294
| # | Model | Provider | Score | Rank | Setting | Source type | Date |
|---|---|---|---|---|---|---|---|
| 1 | Qwen2.5-72BOW | Alibaba (Qwen) | 48.0% | 6 | Average of 6 normalised benchmarks; bfloat16; 72.706B params“Qwen/Qwen2.5-72B-Instruct” | Official leaderboard | Mar 13, 2025 |
| 2 | Qwen2.5-32BOW | Alibaba (Qwen) | 46.6% | 22 | Average of 6 normalised benchmarks; bfloat16; 32.764B params“Qwen/Qwen2.5-32B-Instruct” | Official leaderboard | Mar 13, 2025 |
| 3 | Mistral Large 2 (Nov 2024)OW | Mistral AI | 46.5% | 23 | Average of 6 normalised benchmarks; float16; 122.61B params“mistralai/Mistral-Large-Instruct-2411” | Official leaderboard | Mar 13, 2025 |
| 4 | Llama 3.3 70BOW | Meta | 44.9% | 40 | Average of 6 normalised benchmarks; bfloat16; 70.554B params“meta-llama/Llama-3.3-70B-Instruct” | Official leaderboard | Mar 13, 2025 |
| 5 | Qwen2-72BOW | Alibaba (Qwen) | 43.6% | 58 | Average of 6 normalised benchmarks; bfloat16; 72.706B params“Qwen/Qwen2-72B-Instruct” | Official leaderboard | Mar 13, 2025 |
| 6 | Llama 3.1-70BOW | Meta | 43.4% | 62 | Average of 6 normalised benchmarks; bfloat16; 70.554B params“meta-llama/Llama-3.1-70B-Instruct” | Official leaderboard | Mar 13, 2025 |
| 7 | llama-3.1-tulu-3-70bOW | Ai2 | 42.3% | 103 | Average of 6 normalised benchmarks; bfloat16; 70.554B params“allenai/Llama-3.1-Tulu-3-70B” | Official leaderboard | Mar 13, 2025 |
| 8 | Tulu 3 (Tülu 3) 70BOW | Ai2 | 42.2% | 107 | Average of 6 normalised benchmarks; bfloat16; 70B params“allenai/Llama-3.1-Tulu-3-70B-DPO” | Official leaderboard | Mar 13, 2025 |
| 9 | Qwen2.5-14B-Instruct-1MOW | Alibaba (Qwen) | 41.6% | 148 | Average of 6 normalised benchmarks; bfloat16; 14.77B params“Qwen/Qwen2.5-14B-Instruct-1M” | Official leaderboard | Mar 13, 2025 |
| 10 | Qwen2.5-14BOW | Alibaba (Qwen) | 41.3% | 173 | Average of 6 normalised benchmarks; bfloat16; 14.77B params“Qwen/Qwen2.5-14B-Instruct” | Official leaderboard | Mar 13, 2025 |
| 11 | AceInstruct-72BOW | NVIDIA | 40.4% | 238 | Average of 6 normalised benchmarks; bfloat16; 72.706B params“nvidia/AceInstruct-72B” | Official leaderboard | Mar 13, 2025 |
| 12 | solar-pro-preview-instructOW | upstage | 39.9% | 272 | Average of 6 normalised benchmarks; bfloat16; 22.14B params“upstage/solar-pro-preview-instruct” | Official leaderboard | Mar 13, 2025 |
| 13 | Qwen2.5-Coder-32B-InstructOW | Alibaba (Qwen) | 39.9% | 276 | Average of 6 normalised benchmarks; float16; 32.764B params“Qwen/Qwen2.5-Coder-32B-Instruct” | Official leaderboard | Mar 13, 2025 |
| 14 | Qwen2-VL-72B-InstructOW | Alibaba (Qwen) | 39.5% | 306 | Average of 6 normalised benchmarks; bfloat16; 73.406B params“Qwen/Qwen2-VL-72B-Instruct” | Official leaderboard | Mar 13, 2025 |
| 15 | internlm2_5-20b-chatOW | Shanghai AI Lab (InternLM) | 38.9% | 348 | Average of 6 normalised benchmarks; bfloat16; 19.86B params“internlm/internlm2_5-20b-chat” | Official leaderboard | Mar 13, 2025 |
| 16 | Llama-3.1-Tulu-3-70B-SFTOW | Ai2 | 38.9% | 351 | Average of 6 normalised benchmarks; bfloat16; 70.554B params“allenai/Llama-3.1-Tulu-3-70B-SFT” | Official leaderboard | Mar 13, 2025 |
| 17 | Hermes-3-Llama-3.1-70BOW | NousResearch | 38.5% | 373 | Average of 6 normalised benchmarks; bfloat16; 70.554B params“NousResearch/Hermes-3-Llama-3.1-70B” | Official leaderboard | Mar 13, 2025 |
| 18 | DeepSeek-R1-Distill-Qwen-14BOW | DeepSeek | 38.2% | 392 | Average of 6 normalised benchmarks; bfloat16; 14.77B params“deepseek-ai/DeepSeek-R1-Distill-Qwen-14B” | Official leaderboard | Mar 13, 2025 |
| 19 | Qwen2-Math-72B-InstructOW | Alibaba (Qwen) | 38.0% | 408 | Average of 6 normalised benchmarks; bfloat16; 72.706B params“Qwen/Qwen2-Math-72B-Instruct” | Official leaderboard | Mar 13, 2025 |
| 20 | EXAONE-3.5-32B-InstructOW | LGAI-EXAONE | 37.6% | 435 | Average of 6 normalised benchmarks; float16; 32.003B params“LGAI-EXAONE/EXAONE-3.5-32B-Instruct” | Official leaderboard | Mar 13, 2025 |
| 21 | Llama-3.1-Nemotron-70B-Instruct-HFOW | NVIDIA | 36.9% | 478 | Average of 6 normalised benchmarks; bfloat16; 70.554B params“nvidia/Llama-3.1-Nemotron-70B-Instruct-HF” | Official leaderboard | Mar 13, 2025 |
| 22 | Phi-3.5-MoE-instructOW | Microsoft | 36.9% | 482 | Average of 6 normalised benchmarks; bfloat16; 42B params“microsoft/Phi-3.5-MoE-instruct” | Official leaderboard | Mar 13, 2025 |
| 23 | Qwen2.5-Math-72B-InstructOW | Alibaba (Qwen) | 36.8% | 488 | Average of 6 normalised benchmarks; bfloat16; 72.706B params“Qwen/Qwen2.5-Math-72B-Instruct” | Official leaderboard | Mar 13, 2025 |
| 24 | AceMath-72B-InstructOW | NVIDIA | 36.7% | 504 | Average of 6 normalised benchmarks; bfloat16; 72.706B params“nvidia/AceMath-72B-Instruct” | Official leaderboard | Mar 13, 2025 |
| 25 | Falcon3-7B-InstructOW | TII UAE | 36.4% | 530 | Average of 6 normalised benchmarks; bfloat16; 7.456B params“tiiuae/Falcon3-7B-Instruct” | Official leaderboard | Mar 13, 2025 |
| 26 | Llama 3-70BOW | Meta | 36.4% | 534 | Average of 6 normalised benchmarks; bfloat16; 70.554B params“meta-llama/Meta-Llama-3-70B-Instruct” | Official leaderboard | Mar 13, 2025 |
| 27 | Gemma 2 27BOW | 36.2% | 555 | Average of 6 normalised benchmarks; bfloat16; 27.227B params“google/gemma-2-27b-it” | Official leaderboard | Mar 13, 2025 | |
| 28 | Falcon3-10B-InstructOW | TII UAE | 35.5% | 602 | Average of 6 normalised benchmarks; bfloat16; 10.306B params“tiiuae/Falcon3-10B-Instruct” | Official leaderboard | Mar 13, 2025 |
| 29 | Qwen2.5-7BOW | Alibaba (Qwen) | 35.2% | 623 | Average of 6 normalised benchmarks; bfloat16; 7.616B params“Qwen/Qwen2.5-7B-Instruct” | Official leaderboard | Mar 13, 2025 |
| 30 | QwQ-32B-PreviewOW | Alibaba (Qwen) | 34.1% | 717 | Average of 6 normalised benchmarks; bfloat16; 32.764B params“Qwen/QwQ-32B-Preview” | Official leaderboard | Mar 13, 2025 |
| 31 | Mixtral 8x22BOW | Mistral AI | 33.9% | 731 | Average of 6 normalised benchmarks; bfloat16; 140.621B params“mistralai/Mixtral-8x22B-Instruct-v0.1” | Official leaderboard | Mar 13, 2025 |
| 32 | Command R+OW | Cohere | 33.6% | 748 | Average of 6 normalised benchmarks; float16; 103.811B params“CohereForAI/c4ai-command-r-plus-08-2024” | Official leaderboard | Mar 13, 2025 |
| 33 | Yi-1.5-34BOW | 01.AI | 33.4% | 767 | Average of 6 normalised benchmarks; bfloat16; 34.389B params“01-ai/Yi-1.5-34B-Chat” | Official leaderboard | Mar 13, 2025 |
| 34 | Qwen2.5-Coder-32BOW | Alibaba (Qwen) | 33.3% | 770 | Average of 6 normalised benchmarks; float16; 32.764B params“Qwen/Qwen2.5-Coder-32B” | Official leaderboard | Mar 13, 2025 |
| 35 | Qwen1.5-110BOW | Alibaba (Qwen) | 33.1% | 776 | Average of 6 normalised benchmarks; bfloat16; 111.21B params“Qwen/Qwen1.5-110B-Chat” | Official leaderboard | Mar 13, 2025 |
| 36 | phi-3-medium-4k-instructOW | Microsoft | 33.1% | 778 | Average of 6 normalised benchmarks; bfloat16; 13.96B params“microsoft/Phi-3-medium-4k-instruct” | Official leaderboard | Mar 13, 2025 |
| 37 | AceInstruct-7BOW | NVIDIA | 33.1% | 786 | Average of 6 normalised benchmarks; bfloat16; 7.616B params“nvidia/AceInstruct-7B” | Official leaderboard | Mar 13, 2025 |
| 38 | Qwen2-57B-A14B-InstructOW | Alibaba (Qwen) | 33.0% | 788 | Average of 6 normalised benchmarks; bfloat16; 57.409B params“Qwen/Qwen2-57B-A14B-Instruct” | Official leaderboard | Mar 13, 2025 |
| 39 | internlm2_5-7b-chatOW | Shanghai AI Lab (InternLM) | 33.0% | 792 | Average of 6 normalised benchmarks; float16; 7.738B params“internlm/internlm2_5-7b-chat” | Official leaderboard | Mar 13, 2025 |
| 40 | Qwen2.5-7B-Instruct-1MOW | Alibaba (Qwen) | 32.8% | 798 | Average of 6 normalised benchmarks; bfloat16; 7.616B params“Qwen/Qwen2.5-7B-Instruct-1M” | Official leaderboard | Mar 13, 2025 |
| 41 | EXAONE-3.5-7.8B-InstructOW | LGAI-EXAONE | 32.5% | 809 | Average of 6 normalised benchmarks; float16; 7.818B params“LGAI-EXAONE/EXAONE-3.5-7.8B-Instruct” | Official leaderboard | Mar 13, 2025 |
| 42 | phi-3-small 7.4BOW | Microsoft | 32.3% | 828 | Average of 6 normalised benchmarks; bfloat16; 7.392B params“microsoft/Phi-3-small-8k-instruct” | Official leaderboard | Mar 13, 2025 |
| 43 | Qwen2.5-Coder-14B-InstructOW | Alibaba (Qwen) | 32.1% | 838 | Average of 6 normalised benchmarks; bfloat16; 14.77B params“Qwen/Qwen2.5-Coder-14B-Instruct” | Official leaderboard | Mar 13, 2025 |
| 44 | Gemma 2 9BOW | 32.1% | 843 | Average of 6 normalised benchmarks; bfloat16; 9B params“google/gemma-2-9b-it” | Official leaderboard | Mar 13, 2025 | |
| 45 | phi-3-medium 14BOW | Microsoft | 32.0% | 847 | Average of 6 normalised benchmarks; bfloat16; 13.96B params“microsoft/Phi-3-medium-128k-instruct” | Official leaderboard | Mar 13, 2025 |
| 46 | DeepHermes-3-Mistral-24B-PreviewOW | NousResearch | 32.0% | 853 | Average of 6 normalised benchmarks; bfloat16; 23.572B params“NousResearch/DeepHermes-3-Mistral-24B-Preview” | Official leaderboard | Mar 13, 2025 |
| 47 | Phi-3-small-128k-instructOW | Microsoft | 32.0% | 854 | Average of 6 normalised benchmarks; bfloat16; 7.392B params“microsoft/Phi-3-small-128k-instruct” | Official leaderboard | Mar 13, 2025 |
| 48 | c4ai-command-r7b-12-2024OW | Cohere | 31.6% | 880 | Average of 6 normalised benchmarks; bfloat16; 8.028B params“CohereForAI/c4ai-command-r7b-12-2024” | Official leaderboard | Mar 13, 2025 |
| 49 | c4ai-command-r-plusOW | Cohere | 30.9% | 921 | Average of 6 normalised benchmarks; float16; 103.811B params“CohereForAI/c4ai-command-r-plus” | Official leaderboard | Mar 13, 2025 |
| 50 | Granite-3.2-8B-InstructOW | IBM | 30.8% | 932 | Average of 6 normalised benchmarks; bfloat16; 8.171B params“ibm-granite/granite-3.2-8b-instruct” | Official leaderboard | Mar 13, 2025 |
| 51 | Granite-3.1-8B-InstructOW | IBM | 30.6% | 945 | Average of 6 normalised benchmarks; float16; 8.171B params“ibm-granite/granite-3.1-8b-instruct” | Official leaderboard | Mar 13, 2025 |
| 52 | Phi-4OW | Microsoft Research | 30.4% | 969 | Average of 6 normalised benchmarks; bfloat16; 14.66B params“microsoft/phi-4” | Official leaderboard | Mar 13, 2025 |
| 53 | AceMath-7B-InstructOW | NVIDIA | 30.3% | 972 | Average of 6 normalised benchmarks; bfloat16; 7.616B params“nvidia/AceMath-7B-Instruct” | Official leaderboard | Mar 13, 2025 |
| 54 | Mistral-Small-Instruct-2409OW | Mistral AI | 29.9% | 1044 | Average of 6 normalised benchmarks; float16; 22.247B params“mistralai/Mistral-Small-Instruct-2409” | Official leaderboard | Mar 13, 2025 |
| 55 | Aya-Expanse-32BOW | Cohere | 29.7% | 1075 | Average of 6 normalised benchmarks; float16; 32.296B params“CohereForAI/aya-expanse-32b” | Official leaderboard | Mar 13, 2025 |
| 56 | Yi-1.5-9B-ChatOW | 01.AI | 29.5% | 1105 | Average of 6 normalised benchmarks; bfloat16; 8.829B params“01-ai/Yi-1.5-9B-Chat” | Official leaderboard | Mar 13, 2025 |
| 57 | Phi-4-miniOW | Microsoft | 29.4% | 1121 | Average of 6 normalised benchmarks; bfloat16; 3.836B params“microsoft/Phi-4-mini-instruct” | Official leaderboard | Mar 13, 2025 |
| 58 | Qwen1.5-32BOW | Alibaba (Qwen) | 29.3% | 1145 | Average of 6 normalised benchmarks; bfloat16; 32.512B params“Qwen/Qwen1.5-32B-Chat” | Official leaderboard | Mar 13, 2025 |
| 59 | Phi-3.5-miniOW | Microsoft | 28.2% | 1276 | Average of 6 normalised benchmarks; bfloat16; 3.821B params“microsoft/Phi-3.5-mini-instruct” | Official leaderboard | Mar 13, 2025 |
| 60 | Falcon3-Mamba-7B-InstructOW | TII | 28.1% | 1282 | Average of 6 normalised benchmarks; bfloat16; 7.273B params“tiiuae/Falcon3-Mamba-7B-Instruct” | Official leaderboard | Mar 13, 2025 |
| 61 | Qwen2.5-Coder (7B)OW | Alibaba (Qwen) | 28.1% | 1287 | Average of 6 normalised benchmarks; bfloat16; 7.616B params“Qwen/Qwen2.5-Coder-7B-Instruct” | Official leaderboard | Mar 13, 2025 |
| 62 | Qwen2-7B-InstructOW | Alibaba (Qwen) | 27.9% | 1305 | Average of 6 normalised benchmarks; bfloat16; 7.616B params“Qwen/Qwen2-7B-Instruct” | Official leaderboard | Mar 13, 2025 |
| 63 | DeepSeek-R1-Distill-Llama-70BOW | DeepSeek | 27.8% | 1320 | Average of 6 normalised benchmarks; bfloat16; 70.554B params“deepseek-ai/DeepSeek-R1-Distill-Llama-70B” | Official leaderboard | Mar 13, 2025 |
| 64 | Falcon3-10B-BaseOW | TII | 27.6% | 1334 | Average of 6 normalised benchmarks; bfloat16; 10.306B params“tiiuae/Falcon3-10B-Base” | Official leaderboard | Mar 13, 2025 |
| 65 | phi-3-mini 3.8BOW | Microsoft | 27.6% | 1339 | Average of 6 normalised benchmarks; float16; 3.821B params“microsoft/Phi-3-mini-4k-instruct” | Official leaderboard | Mar 13, 2025 |
| 66 | nous-hermes-2-mixtral-8x7b-dpoOW | NousResearch | 27.4% | 1357 | Average of 6 normalised benchmarks; bfloat16; 46.703B params“NousResearch/Nous-Hermes-2-Mixtral-8x7B-DPO” | Official leaderboard | Mar 13, 2025 |
| 67 | DeepSeek LLM 67BOW | DeepSeek | 27.3% | 1360 | Average of 6 normalised benchmarks; bfloat16; 67B params“deepseek-ai/deepseek-llm-67b-chat” | Official leaderboard | Mar 13, 2025 |
| 68 | Mistral-Small-24B-Base-2501OW | Mistral AI | 27.2% | 1370 | Average of 6 normalised benchmarks; bfloat16; 23.572B params“mistralai/Mistral-Small-24B-Base-2501” | Official leaderboard | Mar 13, 2025 |
| 69 | Qwen2.5-3BOW | Alibaba (Qwen) | 27.2% | 1378 | Average of 6 normalised benchmarks; bfloat16; 3B params“Qwen/Qwen2.5-3B-Instruct” | Official leaderboard | Mar 13, 2025 |
| 70 | EXAONE-3.5-2.4B-InstructOW | LGAI-EXAONE | 27.1% | 1379 | Average of 6 normalised benchmarks; float16; 2.405B params“LGAI-EXAONE/EXAONE-3.5-2.4B-Instruct” | Official leaderboard | Mar 13, 2025 |
| 71 | Falcon3-3B-InstructOW | TII UAE | 26.6% | 1433 | Average of 6 normalised benchmarks; bfloat16; 3.228B params“tiiuae/Falcon3-3B-Instruct” | Official leaderboard | Mar 13, 2025 |
| 72 | qwen2-vl-7b-instructOW | Alibaba (Qwen) | 26.5% | 1443 | Average of 6 normalised benchmarks; bfloat16; 8.291B params“Qwen/Qwen2-VL-7B-Instruct” | Official leaderboard | Mar 13, 2025 |
| 73 | Llama-3.1-Tulu-3-8B-DPOOW | Ai2 | 26.5% | 1446 | Average of 6 normalised benchmarks; bfloat16; 8B params“allenai/Llama-3.1-Tulu-3-8B-DPO” | Official leaderboard | Mar 13, 2025 |
| 74 | llama-3.1-tulu-3-8bOW | Ai2 | 26.3% | 1471 | Average of 6 normalised benchmarks; float16; 8.03B params“allenai/Llama-3.1-Tulu-3-8B” | Official leaderboard | Mar 13, 2025 |
| 75 | c4ai-command-r-v01OW | Cohere | 25.9% | 1516 | Average of 6 normalised benchmarks; float16; 34.981B params“CohereForAI/c4ai-command-r-v01” | Official leaderboard | Mar 13, 2025 |
| 76 | EXAONE-3.0-7.8B-InstructOW | LGAI-EXAONE | 25.7% | 1546 | Average of 6 normalised benchmarks; bfloat16; 7.8B params“LGAI-EXAONE/EXAONE-3.0-7.8B-Instruct” | Official leaderboard | Mar 13, 2025 |
| 77 | DBRXOW | Databricks | 25.2% | 1607 | Average of 6 normalised benchmarks; bfloat16; 131.597B params“databricks/dbrx-instruct” | Official leaderboard | Mar 13, 2025 |
| 78 | Qwen2.5-Coder-14BOW | Alibaba (Qwen) | 24.8% | 1651 | Average of 6 normalised benchmarks; bfloat16; 14.77B params“Qwen/Qwen2.5-Coder-14B” | Official leaderboard | Mar 13, 2025 |
| 79 | Hermes-2-Theta-Llama-3-8BOW | NousResearch | 24.8% | 1657 | Average of 6 normalised benchmarks; bfloat16; 8.03B params“NousResearch/Hermes-2-Theta-Llama-3-8B” | Official leaderboard | Mar 13, 2025 |
| 80 | aya-23-35BOW | Cohere | 24.8% | 1661 | Average of 6 normalised benchmarks; float16; 34.981B params“CohereForAI/aya-23-35B” | Official leaderboard | Mar 13, 2025 |
| 81 | Falcon3-7B-BaseOW | TII | 24.8% | 1665 | Average of 6 normalised benchmarks; bfloat16; 7.456B params“tiiuae/Falcon3-7B-Base” | Official leaderboard | Mar 13, 2025 |
| 82 | Mistral NeMoOW | Mistral AI | 24.7% | 1676 | Average of 6 normalised benchmarks; bfloat16; 12.248B params“mistralai/Mistral-Nemo-Instruct-2407” | Official leaderboard | Mar 13, 2025 |
| 83 | Yi-34BOW | 01.AI | 24.2% | 1759 | Average of 6 normalised benchmarks; bfloat16; 34.389B params“01-ai/Yi-34B-Chat” | Official leaderboard | Mar 13, 2025 |
| 84 | Llama-3.2-3B-InstructOW | Meta | 24.2% | 1768 | Average of 6 normalised benchmarks; bfloat16; 3.213B params“meta-llama/Llama-3.2-3B-Instruct” | Official leaderboard | Mar 13, 2025 |
| 85 | Ministral 8BOW | Mistral AI | 24.2% | 1776 | Average of 6 normalised benchmarks; bfloat16; 8.02B params“mistralai/Ministral-8B-Instruct-2410” | Official leaderboard | Mar 13, 2025 |
| 86 | granite-3.0-8b-instructOW | IBM | 24.0% | 1825 | Average of 6 normalised benchmarks; bfloat16; 8.171B params“ibm-granite/granite-3.0-8b-instruct” | Official leaderboard | Mar 13, 2025 |
| 87 | Llama 3-8BOW | Meta | 23.9% | 1868 | Average of 6 normalised benchmarks; bfloat16; 8.03B params“meta-llama/Meta-Llama-3-8B-Instruct” | Official leaderboard | Mar 13, 2025 |
| 88 | Mixtral 8x7BOW | Mistral AI | 23.8% | 1884 | Average of 6 normalised benchmarks; bfloat16; 46.703B params“mistralai/Mixtral-8x7B-Instruct-v0.1” | Official leaderboard | Mar 13, 2025 |
| 89 | Llama 3.1-8BOW | Meta | 23.8% | 1905 | Average of 6 normalised benchmarks; float16; 8.03B params“meta-llama/Llama-3.1-8B-Instruct” | Official leaderboard | Mar 13, 2025 |
| 90 | Mistral-NeMo-Minitron-8B-InstructOW | NVIDIA | 23.6% | 1949 | Average of 6 normalised benchmarks; bfloat16; 8.414B params“nvidia/Mistral-NeMo-Minitron-8B-Instruct” | Official leaderboard | Mar 13, 2025 |
| 91 | Qwen1.5-14BOW | Alibaba (Qwen) | 23.6% | 1952 | Average of 6 normalised benchmarks; bfloat16; 14.167B params“Qwen/Qwen1.5-14B-Chat” | Official leaderboard | Mar 13, 2025 |
| 92 | Hermes-3-Llama-3.1-8BOW | NousResearch | 23.5% | 1976 | Average of 6 normalised benchmarks; bfloat16; 8.03B params“NousResearch/Hermes-3-Llama-3.1-8B” | Official leaderboard | Mar 13, 2025 |
| 93 | Nous-Hermes-2-SOLAR-10.7BOW | NousResearch | 23.4% | 2003 | Average of 6 normalised benchmarks; bfloat16; 10.732B params“NousResearch/Nous-Hermes-2-SOLAR-10.7B” | Official leaderboard | Mar 13, 2025 |
| 94 | Codestral-22B-v0.1OW | Mistral AI | 23.3% | 2028 | Average of 6 normalised benchmarks; bfloat16; 22.247B params“mistralai/Codestral-22B-v0.1” | Official leaderboard | Mar 13, 2025 |
| 95 | DeepSeek-R1-Distill-Qwen-32BOW | DeepSeek | 23.0% | 2100 | Average of 6 normalised benchmarks; bfloat16; 32.764B params“deepseek-ai/DeepSeek-R1-Distill-Qwen-32B” | Official leaderboard | Mar 13, 2025 |
| 96 | Stable Beluga 2OW | Stability AI | 22.8% | 2129 | Average of 6 normalised benchmarks; bfloat16; 68.977B params“stabilityai/StableBeluga2” | Official leaderboard | Mar 13, 2025 |
| 97 | Yi-1.5-6B-ChatOW | 01.AI | 22.8% | 2136 | Average of 6 normalised benchmarks; bfloat16; 6.061B params“01-ai/Yi-1.5-6B-Chat” | Official leaderboard | Mar 13, 2025 |
| 98 | Llama-3.1-Tulu-3-8B-SFTOW | Ai2 | 22.6% | 2187 | Average of 6 normalised benchmarks; bfloat16; 8.03B params“allenai/Llama-3.1-Tulu-3-8B-SFT” | Official leaderboard | Mar 13, 2025 |
| 99 | Aya-Expanse-8BOW | Cohere | 22.4% | 2218 | Average of 6 normalised benchmarks; float16; 8.028B params“CohereForAI/aya-expanse-8b” | Official leaderboard | Mar 13, 2025 |
| 100 | Yi-1.5-9BOW | 01.AI | 22.1% | 2256 | Average of 6 normalised benchmarks; bfloat16; 8.829B params“01-ai/Yi-1.5-9B” | Official leaderboard | Mar 13, 2025 |
Data sources & licences
| Source | Licence / terms | How we reuse it | Last fetched |
|---|---|---|---|
| Hugging Face Open LLM Leaderboard v2 (archived) Hugging Face, Open LLM Leaderboard (archived 2025) | No licence tag – factual scores only, with citation | No explicit licence – factual scores only, with citation 4576 rows in archive; 353 official-developer models kept | Oct 22, 2026 |
FAQ
What does Open LLM Leaderboard v2 measure?
Hugging Face's leaderboard of open-weight models, scored on IFEval, BBH, MATH Lvl 5, GPQA, MuSR and MMLU-Pro with normalised scores. The leaderboard was archived in 2025 and no longer accepts submissions. Average of six normalised academic benchmarks, run by Hugging Face with the same harness.
How reliable is Open LLM Leaderboard v2?
Archived: no models after early 2025. Only open-weight models. We keep only models published by their original developers (no merges or community fine-tunes). Contamination risk: High – Fixed public test sets; the leaderboard itself flagged contaminated submissions.