Benchmark
MMLU
Scores are percentages. Higher is better.
- Models and systems measured
- 138
- Scores recorded
- 153
- Sources
- 3
- IQ weight
- 0.00
Score distribution
- 76.9%
- Mean
- 80.1%
- Median
- 12.2%
- Std dev
- 60.1%
- 10th percentile
- 88.8%
- 90th percentile
- 36.6%
- Min
This benchmark has a weight of 0.00 for chat models, measured over 133 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.
Leaderboard
| # | Model or system | Score |
|---|---|---|
| 1 | GPT 5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 92.5% |
| 2 | Doubao Seed 1.8· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 92.3% |
| 3 | o1· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 91.8% |
| 4 | GPT-4.5 Preview· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 90.8% |
| 5 | O1 Preview· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 90.8% |
| 6 | Qwen3 VL 235Bthinking· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. | 90.6% |
| 7 | Sarvam 105B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 90.6% |
| 8 | Claude 3.5 Sonnet· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claimValues from 2 sources
| 90.4% |
| 9 | GPT 4.1· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 90.2% |
| 10 | Kimi K2 0905· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 90.2% |
| 11 | GPT OSS 120B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 90% |
| 12 | GPT OSS 120Bofficial model card Table 3 · reasoning high · published setting; Table 3 does not further specify tool access · pass@1· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 90% |
| 13 | LongCat Flash Chat· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 89.7% |
| 14 | Kimi K2 Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 89.5% |
| 15 | MiMo V2.5 Pro· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 89.4% |
| 16 | Qwen3 VL 235B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. | 88.8% |
| 17 | Qwen3 VL 32B Thinking· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. | 88.7% |
| 18 | DeepSeek V3· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claimValues from 2 sources
| 88.5% |
| 19 | GPT 4o· Epoch AI Benchmarking Hub dataValues from 2 sources
| 88.1% |
| 20 | GPT OSS 120Bofficial model card Table 3 · reasoning medium · published setting; Table 3 does not further specify tool access · pass@1· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 88% |
| 21 | Qwen 3 235B A22B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 87.8% |
| 22 | qwen3-vl-30bthinking· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. | 87.6% |
| 23 | GPT-4.1 mini· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 87.5% |
| 24 | Grok 2· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 87.5% |
| 25 | Llama 3.1 405B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claimValues from 2 sources
| 87.3% |
| 26 | Trinity Large PreviewTrinity technical report · Table 4 · exact reported precision· Reported by Arcee AI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 87.2% |
| 27 | Trinity Large PreviewTrinity-Large-Preview Model Card · rounded to one decimal· Reported by Arcee AI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 87.2% |
| 28 | gemini-1.5-pro-002· Epoch AI Benchmarking Hub data | 86.9% |
| 29 | o3 Mini· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 86.9% |
| 30 | Claude 3 Opus· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claimValues from 2 sources
| 86.8% |
| 31 | GPT-4 Turbo· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claimValues from 2 sources
| 86.5% |
| 32 | GPT 4· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claimValues from 2 sources
| 86.4% |
| 33 | Qwen3 VL 32B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. | 86.4% |
| 34 | Llama 3.3 70B Instruct· Epoch AI Benchmarking Hub dataValues from 2 sources
| 86.3% |
| 35 | Grok 2 mini· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 86.2% |
| 36 | Particle 1· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 86.1% |
| 37 | Llama 3.1 70B Instruct0-shot · chain of thought· Reported by Meta· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 86% |
| 38 | gemini-1.5-pro-001· Epoch AI Benchmarking Hub data | 85.9% |
| 39 | GPT OSS 120Bofficial model card Table 3 · reasoning low · published setting; Table 3 does not further specify tool access · pass@1· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 85.9% |
| 40 | Gemini 1.5 Pro· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 85.9% |
| 41 | Nova Pro· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 85.9% |
| 42 | longcat-flash-lite· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 85.5% |
| 43 | Llama 4 Maverick· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 85.5% |
| 44 | GPT OSS 20b· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 85.3% |
| 45 | GPT OSS 20bofficial model card Table 3 · reasoning high · published setting; Table 3 does not further specify tool access · pass@1· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 85.3% |
| 46 | Qwen2.5 72B Instruct· Epoch AI Benchmarking Hub data | 85.3% |
| 47 | O1 Mini· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 85.2% |
| 48 | Qwen: Qwen3 VL 8B Thinking· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. | 85.2% |
| 49 | Sarvam-30B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 85.1% |
| 50 | Qwen2.5 72B· Epoch AI Benchmarking Hub data | 85% |
Showing 1–50 of 148 results.
Not ranked
The source marks these results as descriptive, not capability scores, so we list them by name instead of ranking them.
| Model or system | Score |
|---|---|
| MiMo V2 Flash Baseofficial Base table · 5-shot· Reported by Xiaomi MiMo· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 86.7% |
| Qwen3.5-397B-A17B-BaseQwen3.5 release/model card · base-model table· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 88.6% |
| Qwen3.7-Plus-BaseQwen3.8-Flash-Next release/model card · base-model table· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 90.4% |
| Qwen3.8-27B-BaseQwen3.8-Flash-Next release/model card · base-model table· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 87.5% |
| Qwen3.8-Flash-Next-BaseQwen3.8-Flash-Next release/model card · base-model table· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 90.4% |
Data credits: Epoch AI Benchmarking Hub, CC BY 4.0 (licence) · Data by LLM Stats · BenchBoard by SYuan03 (github.com/SYuan03/benchboard). MIT covers its code and data organization; no results licence is inferred. Numeric relays retain the original stating-source links.