Benchmark
GPQA Diamond
Science benchmarks measure how well models answer science questions. Scores are percentages. Higher is better.
- Models and systems measured
- 282
- Scores recorded
- 1623
- Sources
- 4
- IQ weight
- 1.72
Score distribution
- 79.3%
- Mean
- 84.5%
- Median
- 15.7%
- Std dev
- 56.8%
- 10th percentile
- 92.2%
- 90th percentile
- 0.168%
- Min
This benchmark has a weight of 1.72 for chat models, measured over 272 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 | |
|---|---|---|
| 1501 | GLM 4.7 FlashOpenRouter harness, z-ai/glm-4.7-flash-20260119, Venice, public 90-day aggregate· OpenRouter benchmark exports | 49.8% |
| 1502 | Llama 3.3 70B InstructOpenRouter harness, meta-llama/llama-3.3-70b-instruct, Groq, public 90-day aggregate· OpenRouter benchmark exports | 49.5% |
| 1503 | Llama 3.3 70B InstructOpenRouter harness, meta-llama/llama-3.3-70b-instruct, Cloudflare, public 90-day aggregate· OpenRouter benchmark exports | 49.3% |
| 1504 | Qwen3 14BOpenRouter harness, qwen/qwen3-14b-04-28, Alibaba, public 90-day aggregate· OpenRouter benchmark exports | 49.2% |
| 1505 | Qwen2.5 72B Instruct· Epoch AI Benchmarking Hub data | 49.1% |
| 1506 | GPT-4.1 nanoOpenRouter harness, openai/gpt-4.1-nano-2025-04-14, auto-routing, public 90-day aggregate· OpenRouter benchmark exports | 49.1% |
| 1507 | Mistral Small 3.2· Epoch AI Benchmarking Hub data | 49.1% |
| 1508 | Mistral Large2407· Epoch AI Benchmarking Hub data | 49% |
| 1509 | GPT-4.1 nano· Epoch AI Benchmarking Hub data | 48.9% |
| 1510 | Llama 3.3 70B InstructOpenRouter harness, meta-llama/llama-3.3-70b-instruct, CoreWeave, public 90-day aggregate· OpenRouter benchmark exports | 48.7% |
| 1511 | GLM 4.7 FlashOpenRouter harness, z-ai/glm-4.7-flash-20260119, DeepInfra, public 90-day aggregate· OpenRouter benchmark exports | 48.7% |
| 1512 | GPT-5 nanominimal· Epoch AI Benchmarking Hub data | 48.5% |
| 1513 | Llama 3.3 70B InstructOpenRouter harness, meta-llama/llama-3.3-70b-instruct, AkashML, public 90-day aggregate· OpenRouter benchmark exports | 48.3% |
| 1514 | Llama 3.3 70B InstructOpenRouter harness, meta-llama/llama-3.3-70b-instruct, auto-routing, public 90-day aggregate· OpenRouter benchmark exports | 48.2% |
| 1515 | Llama 3.3 70B InstructOpenRouter harness, meta-llama/llama-3.3-70b-instruct, SambaNova, public 90-day aggregate· OpenRouter benchmark exports | 48.1% |
| 1516 | Qwen Plus· Epoch AI Benchmarking Hub data | 48.1% |
| 1517 | GPT 4o· Epoch AI Benchmarking Hub data | 47.9% |
| 1518 | Llama 3.3 70B InstructOpenRouter harness, meta-llama/llama-3.3-70b-instruct, Parasail, public 90-day aggregate· OpenRouter benchmark exports | 47.9% |
| 1519 | Gemma 3 27B IT· Epoch AI Benchmarking Hub data | 47.7% |
| 1520 | Llama 3.3 70B InstructOpenRouter harness, meta-llama/llama-3.3-70b-instruct, Together, public 90-day aggregate· OpenRouter benchmark exports | 47.7% |
| 1521 | Qwen2.5 72B InstructOpenRouter harness, qwen/qwen-2.5-72b-instruct, DeepInfra, public 90-day aggregate· OpenRouter benchmark exports | 47.7% |
| 1522 | Magistral Small 2509· Epoch AI Benchmarking Hub data | 47.6% |
| 1523 | Mistral Small2503· Epoch AI Benchmarking Hub data | 47.5% |
| 1524 | Qwen3 8Bnone· Epoch AI Benchmarking Hub data | 47.5% |
| 1525 | Llama 3.3 70B Instruct· Epoch AI Benchmarking Hub data | 47.4% |
| 1526 | gemini-1.5-flash-002· Epoch AI Benchmarking Hub data | 47.3% |
| 1527 | Mistral-Small-3· Epoch AI Benchmarking Hub data | 47.3% |
| 1528 | Llama 3.3 70B InstructOpenRouter harness, meta-llama/llama-3.3-70b-instruct, Novita, public 90-day aggregate· OpenRouter benchmark exports | 47.2% |
| 1529 | Claude 3 Opus· Epoch AI Benchmarking Hub data | 47.2% |
| 1530 | Llama 3.3 70B InstructOpenRouter harness, meta-llama/llama-3.3-70b-instruct, PrimeIntellect, public 90-day aggregate· OpenRouter benchmark exports | 47% |
| 1531 | Llama 3.3 70B InstructOpenRouter harness, meta-llama/llama-3.3-70b-instruct, Crusoe, public 90-day aggregate· OpenRouter benchmark exports | 46.9% |
| 1532 | GPT-4 Turbo· Epoch AI Benchmarking Hub data | 46.6% |
| 1533 | Llama 3.3 70B InstructOpenRouter harness, meta-llama/llama-3.3-70b-instruct, DeepInfra, public 90-day aggregate· OpenRouter benchmark exports | 46.3% |
| 1534 | Qwen2.5 72B InstructOpenRouter harness, qwen/qwen-2.5-72b-instruct, Novita, public 90-day aggregate· OpenRouter benchmark exports | 46.1% |
| 1535 | Qwen2.5 32B Instruct· Epoch AI Benchmarking Hub data | 46.1% |
| 1536 | Qwen2.5 72B InstructOpenRouter harness, qwen/qwen-2.5-72b-instruct, auto-routing, public 90-day aggregate· OpenRouter benchmark exports | 46% |
| 1537 | DeepSeek V3OpenRouter harness, deepseek/deepseek-chat-v3-0324, auto-routing, public 90-day aggregate· OpenRouter benchmark exports | 46% |
| 1538 | GPT OSS 20bhigh· Epoch AI Benchmarking Hub data | 46% |
| 1539 | GPT OSS 120BOpenRouter harness, openai/gpt-oss-120b, DigitalOcean, public 90-day aggregate· OpenRouter benchmark exports | 45.9% |
| 1540 | gemini-1.5-pro-001· Epoch AI Benchmarking Hub data | 45.9% |
| 1541 | Qwen2.5 VL 72B InstructOpenRouter harness, qwen/qwen2.5-vl-72b-instruct, Parasail, public 90-day aggregate· OpenRouter benchmark exports | 45.7% |
| 1542 | Llama 3.3 70B InstructOpenRouter harness, meta-llama/llama-3.3-70b-instruct, Nebius, public 90-day aggregate· OpenRouter benchmark exports | 45.6% |
| 1543 | Mistral Small· Epoch AI Benchmarking Hub data | 45.3% |
| 1544 | GLM 4.7 Flashnone· Epoch AI Benchmarking Hub data | 45.1% |
| 1545 | DeepSeek R1 Distill Qwen 14B· Epoch AI Benchmarking Hub data | 44.7% |
| 1546 | Qwen2.5 VL 72B InstructOpenRouter harness, qwen/qwen2.5-vl-72b-instruct, Nebius, public 90-day aggregate· OpenRouter benchmark exports | 44.3% |
| 1547 | Llama 3.1 70B Instruct· Epoch AI Benchmarking Hub data | 44.2% |
| 1548 | GPT-4o miniOpenRouter harness, openai/gpt-4o-mini, auto-routing, public 90-day aggregate· OpenRouter benchmark exports | 43.6% |
| 1549 | Qwen3 4Bnone· Epoch AI Benchmarking Hub data | 43.6% |
| 1550 | WizardLM 2 8x22B· Epoch AI Benchmarking Hub data | 43.4% |
Showing 1,501–1,550 of 1,621 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 | |
|---|---|
| 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. | 55.1% |
| Qwen3.8-2.4T-A95Blifecycle-page core-benchmark claim· Reported by Alibaba Cloud Model Studio· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 92.6% |
Tokens the model generated per task, as measured by the source. Comparable within this benchmark only.
Data credits: Epoch AI Benchmarking Hub, CC BY 4.0 (licence) · 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. · OpenRouter public benchmark leaderboard and search-lane exports, CC BY 4.0. Fractional source scores converted to percentages; immutable dated snapshots. (licence) · OpenEvals/leaderboard-data, MIT (licence)