Skip to content

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
0.168%96.8%
The band covers the middle 80% of scores. The line marks the median.

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
1101Qwen 3.8 27BOpenRouter harness, qwen/qwen3.8-27b-20260814, Venice, public 90-day aggregate· OpenRouter benchmark exports79.1%
1102DeepSeek V3.1OpenRouter harness, deepseek/deepseek-chat-v3.1, CoreWeave, public 90-day aggregate· OpenRouter benchmark exports79%
1103Gemma 4 26B A4BOpenRouter harness, google/gemma-4-26b-a4b-it-20260403, Google, public 90-day aggregate· OpenRouter benchmark exports79%
1104Kimi K2.5OpenRouter harness, moonshotai/kimi-k2.5-0127, AtlasCloud, public 90-day aggregate· OpenRouter benchmark exports79%
1105DeepSeek V3.2OpenRouter harness, deepseek/deepseek-v3.2-20251201, Novita, public 90-day aggregate· OpenRouter benchmark exports79%
1106Qwen3.5 9Bnone· Epoch AI Benchmarking Hub data78.9%
1107Claude Fable 5low· Epoch AI Benchmarking Hub data78.8%
1108Claude Sonnet 4.516k· Epoch AI Benchmarking Hub data78.8%
1109Claude Sonnet 4.6max· Epoch AI Benchmarking Hub data78.8%
1110DeepSeek V4 Flash 0731OpenRouter harness, deepseek/deepseek-v4-flash-20260731, OpenInference, public 90-day aggregate· OpenRouter benchmark exports78.8%
1111MiMo V2.5OpenRouter harness, xiaomi/mimo-v2.5-20260422, auto-routing, public 90-day aggregate· OpenRouter benchmark exports78.8%
1112GLM 4.6OpenRouter harness, z-ai/glm-4.6, DeepInfra, public 90-day aggregate· OpenRouter benchmark exports78.7%
1113GLM 5.2OpenRouter harness, z-ai/glm-5.2-20260616, Phala, public 90-day aggregate· OpenRouter benchmark exports78.7%
1114Gemma 4 26B A4BOpenRouter harness, google/gemma-4-26b-a4b-it-20260403, PrimeIntellect, public 90-day aggregate· OpenRouter benchmark exports78.7%
1115DeepSeek V3.2OpenRouter harness, deepseek/deepseek-v3.2-20251201, DeepInfra, public 90-day aggregate· OpenRouter benchmark exports78.7%
1116Gemma 4 26B A4BOpenRouter harness, google/gemma-4-26b-a4b-it-20260403, Io Net, public 90-day aggregate· OpenRouter benchmark exports78.7%
1117Claude Sonnet 4OpenRouter harness, anthropic/claude-4-sonnet-20250522, auto-routing, public 90-day aggregate· OpenRouter benchmark exports78.6%
1118DeepSeek R1OpenRouter harness, deepseek/deepseek-r1-0528, DeepInfra, public 90-day aggregate· OpenRouter benchmark exports78.6%
1119MiniMax M2.7OpenRouter harness, minimax/minimax-m2.7-20260318, Mara, public 90-day aggregate· OpenRouter benchmark exports78.5%
1120Qwen3.5 35B A3BOpenRouter harness, qwen/qwen3.5-35b-a3b-20260224, Darkbloom, public 90-day aggregate· OpenRouter benchmark exports78.5%
1121Claude Sonnet 3.764k· Epoch AI Benchmarking Hub data78.5%
1122GLM 5.3 FlashOpenRouter harness, z-ai/glm-5.3-flash-20260826, Nebius, public 90-day aggregate· OpenRouter benchmark exports78.5%
1123GPT 5.4 Nanohigh· Epoch AI Benchmarking Hub data78.5%
1124DeepSeek V3.1 TerminusOpenRouter harness, deepseek/deepseek-v3.1-terminus, StreamLake, public 90-day aggregate· OpenRouter benchmark exports78.5%
1125MiniMax M2.5OpenRouter harness, minimax/minimax-m2.5-20260211, AtlasCloud, public 90-day aggregate· OpenRouter benchmark exports78.5%
1126MiMo V2.5 ProOpenRouter harness, xiaomi/mimo-v2.5-pro-20260422, GMICloud, public 90-day aggregate· OpenRouter benchmark exports78.4%
1127Qwen3.5 35B A3BOpenRouter harness, qwen/qwen3.5-35b-a3b-20260224, AtlasCloud, public 90-day aggregate· OpenRouter benchmark exports78.4%
1128Step 3.7 FlashStep 3.7 Flash official README/figure · avg@16 · NVFP4 · MTP disabled· Reported by StepFun· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.78.4%
1129MiMo V2.6 FlashOpenRouter harness, xiaomi/mimo-v2.6-flash-20260921, GMICloud, public 90-day aggregate· OpenRouter benchmark exports78.3%
1130GLM 5.1OpenRouter harness, z-ai/glm-5.1-20260406, AtlasCloud, public 90-day aggregate· OpenRouter benchmark exports78.3%
1131Claude Sonnet 432k· Epoch AI Benchmarking Hub data78.3%
1132GPT OSS 120BOpenRouter harness, openai/gpt-oss-120b, PrimeIntellect, public 90-day aggregate· OpenRouter benchmark exports78.2%
1133Qwen3.6 35B A3BOpenRouter harness, qwen/qwen3.6-35b-a3b-20260415, PrimeIntellect, public 90-day aggregate· OpenRouter benchmark exports77.9%
1134Claude Sonnet 459k· Epoch AI Benchmarking Hub data77.8%
1135Step 3.7 FlashStep 3.7 Flash official README/figure · avg@16 · NVFP4 + MTP· Reported by StepFun· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.77.8%
1136o4 Minimedium· Epoch AI Benchmarking Hub data77.8%
1137GPT 5.4 NanoOpenRouter harness, openai/gpt-5.4-nano-20260317, auto-routing, public 90-day aggregate· OpenRouter benchmark exports77.7%
1138GPT OSS 120BOpenRouter harness, openai/gpt-oss-120b, CoreWeave, public 90-day aggregate· OpenRouter benchmark exports77.6%
1139GPT 5.4 NanoOpenRouter harness, openai/gpt-5.4-nano-20260317, OpenAI, public 90-day aggregate· OpenRouter benchmark exports77.6%
1140Doubao-Seed 1.6· Epoch AI Benchmarking Hub data77.6%
1141GLM 4.6OpenRouter harness, z-ai/glm-4.6, auto-routing, public 90-day aggregate· OpenRouter benchmark exports77.6%
1142GPT OSS 120BOpenRouter harness, openai/gpt-oss-120b, Parasail, public 90-day aggregate· OpenRouter benchmark exports77.5%
1143DeepSeek V3.1OpenRouter harness, deepseek/deepseek-chat-v3.1, DeepInfra, public 90-day aggregate· OpenRouter benchmark exports77.4%
1144MiMo V2.6 FlashOpenRouter harness, xiaomi/mimo-v2.6-flash-20260921, DeepInfra, public 90-day aggregate· OpenRouter benchmark exports77.4%
1145Qwen3.5 9BOpenRouter harness, qwen/qwen3.5-9b-20260310, auto-routing, public 90-day aggregate· OpenRouter benchmark exports77.4%
1146GPT OSS 120BOpenRouter harness, openai/gpt-oss-120b, DekaLLM, public 90-day aggregate· OpenRouter benchmark exports77.4%
1147Gemma 4 31B ITOpenRouter harness, google/gemma-4-31b-it-20260402, Cerebras, public 90-day aggregate· OpenRouter benchmark exports77.4%
1148GPT 5.4 NanoOpenRouter harness, openai/gpt-5.4-nano-20260317, Azure, public 90-day aggregate· OpenRouter benchmark exports77.4%
1149MiMo V2.5OpenRouter harness, xiaomi/mimo-v2.5-20260422, GMICloud, public 90-day aggregate· OpenRouter benchmark exports77.3%
1150Claude Opus 4.116k· Epoch AI Benchmarking Hub data77.3%

Showing 1,101–1,150 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)