Benchmark
GPQA
Science benchmarks measure how well models answer science questions. Scores are percentages. Higher is better.
- Models and systems measured
- 364
- Scores recorded
- 378
- Sources
- 3
- IQ weight
- 0.75
Score distribution
- 45.3%
- Mean
- 47.5%
- Median
- 34.4%
- Std dev
- 2.9%
- 10th percentile
- 89%
- 90th percentile
- 0%
- Min
This benchmark has a weight of 0.75 for chat models, measured over 359 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-6.2 Astra· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 97% |
| 2 | GPT-6 Astra· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 96% |
| 3 | GPT-6.2 Sol· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 96% |
| 4 | GPT-5.6 Sol· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 94.6% |
| 5 | Gemini 3.1 Pro Preview· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 94.3% |
| 6 | Claude Opus 4.7· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 94.2% |
| 7 | Kimi K3.1 Pro· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 94.2% |
| 8 | Claude Opus 4.8· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 93.6% |
| 9 | GPT 5.5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 93.6% |
| 10 | Kimi K3· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 93.5% |
| 11 | GPT-5.2 Pro· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 93.2% |
| 12 | Grok 4.5· LLM Stats data | 93% |
| 13 | Inkling 2 Small· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 93% |
| 14 | GPT-5.6 Terra· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 92.9% |
| 15 | GPT 5.4· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 92.8% |
| 16 | Qwen3.8 Max· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 92.6% |
| 17 | GPT 5.2· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 92.4% |
| 18 | Qwen3.7 Max· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 92.4% |
| 19 | GPT-5.6 Luna· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 92.3% |
| 20 | Hy4 preview· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 92.3% |
| 21 | Gemini 3 Pro Preview· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 91.9% |
| 22 | Qwen 3.8 Flash Next· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 91.7% |
| 23 | Qwen3.8 Flash· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 91.7% |
| 24 | Claude Opus 4.6· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 91.3% |
| 25 | GLM 5.2· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 91.2% |
| 26 | DeepSeek V4.1 Flash· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 90.9% |
| 27 | Beam 1.0 501B A23B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 90.6% |
| 28 | Kimi K2.6· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 90.5% |
| 29 | Gemini 3 Flash· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 90.4% |
| 30 | Hy3 (Deep Infra)· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 90.4% |
| 31 | Particle 1· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 90.4% |
| 32 | Qwen3.6 Plus· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 90.4% |
| 33 | Qwen3.7 Plus· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 90.3% |
| 34 | Claude Sonnet 4.6· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 89.9% |
| 35 | NVIDIA Nemotron 3.5 Super VL 120B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 89.5% |
| 36 | Inkling Small· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 89.5% |
| 37 | Muse Spark· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 89.5% |
| 38 | Mercury 3· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 89.2% |
| 39 | Qwen 3.8 27B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 89.2% |
| 40 | Muse Spark Open· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 89% |
| 41 | Solar Pro 4· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 89% |
| 42 | Doubao Seed 2.0 Pro· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 88.9% |
| 43 | Qwen3.5 397B A17B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 88.4% |
| 44 | Qwen3.5 397B A17BQwen3.5 release/model card · published setting· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 88.4% |
| 45 | GPT 5.1· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 88.1% |
| 46 | GPT 5.1high· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 88.1% |
| 47 | GPT-5.1-Instant· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 88.1% |
| 48 | GPT-5.4 mini· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 88% |
| 49 | Qwen3.6 27B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 87.8% |
| 50 | Kimi K2.5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 87.6% |
Showing 1–50 of 374 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 |
|---|---|
| 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. | 54.6% |
| 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. | 51.5% |
| 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. | 45% |
| 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. | 51.4% |
Data credits: 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.