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
751GPT 5medium· Epoch AI Benchmarking Hub data85.4%
752Nemotron 3 Ultra· Epoch AI Benchmarking Hub data85.4%
753GLM 5.2OpenRouter harness, z-ai/glm-5.2-20260616, Sail Research, public 90-day aggregate· OpenRouter benchmark exports85.4%
754GLM 5.3OpenRouter harness, z-ai/glm-5.3-20260816, Inceptron, public 90-day aggregate· OpenRouter benchmark exports85.3%
755GLM 5.3OpenRouter harness, z-ai/glm-5.3-20260816, Baidu, public 90-day aggregate· OpenRouter benchmark exports85.3%
756GLM 5.3 FlashOpenRouter harness, z-ai/glm-5.3-flash-20260826, GMICloud, public 90-day aggregate· OpenRouter benchmark exports85.3%
757Gemma 4 31B ITOpenRouter harness, google/gemma-4-31b-it-20260402, Parasail, public 90-day aggregate· OpenRouter benchmark exports85.3%
758Gemini 2.5 Pro· Epoch AI Benchmarking Hub data85.3%
759Qwen3.6 35B A3BOpenRouter harness, qwen/qwen3.6-35b-a3b-20260415, Parasail, public 90-day aggregate· OpenRouter benchmark exports85.3%
760GLM 5.3OpenRouter harness, z-ai/glm-5.3-20260816, DekaLLM, public 90-day aggregate· OpenRouter benchmark exports85.2%
761MiMo V2.6 FlashOpenRouter harness, xiaomi/mimo-v2.6-flash-20260921, Venice, public 90-day aggregate· OpenRouter benchmark exports85.2%
762Gemma 4 31B ITOpenRouter harness, google/gemma-4-31b-it-20260402, Together, public 90-day aggregate· OpenRouter benchmark exports85.2%
763MiniMax M2.5· OpenEvals Leaderboard data· Reported by the model's developer. The source relaying it did not verify it.85.2%
764MiniMax M2.5official appendix· Reported by MiniMax· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.85.2%
765MiniMax M3.1 FlashOpenRouter harness, minimax/minimax-m3.1-flash-20261019, Tenstorrent, public 90-day aggregate· OpenRouter benchmark exports85.2%
766GPT 5OpenRouter harness, openai/gpt-5-2025-08-07, Azure, public 90-day aggregate· OpenRouter benchmark exports85.2%
767GLM 5.2OpenRouter harness, z-ai/glm-5.2-20260616, Alibaba, public 90-day aggregate· OpenRouter benchmark exports85.2%
768GLM 5.2OpenRouter harness, z-ai/glm-5.2-20260616, auto-routing, public 90-day aggregate· OpenRouter benchmark exports85.1%
769MiniMax M3.1 FlashOpenRouter harness, minimax/minimax-m3.1-flash-20261019, Parasail, public 90-day aggregate· OpenRouter benchmark exports85.1%
770Qwen3.5 122B-A10BOpenRouter harness, qwen/qwen3.5-122b-a10b-20260224, auto-routing, public 90-day aggregate· OpenRouter benchmark exports85.1%
771GLM 5.1OpenRouter harness, z-ai/glm-5.1-20260406, StreamLake, public 90-day aggregate· OpenRouter benchmark exports85%
772GPT 5.1medium· Epoch AI Benchmarking Hub data85%
773Nemotron 3.5 Super VLOpenRouter harness, nvidia/nemotron-3.5-super-vl-20261015, Darkbloom, public 90-day aggregate· OpenRouter benchmark exports85%
774GPT 5OpenRouter harness, openai/gpt-5-2025-08-07, OpenAI, public 90-day aggregate· OpenRouter benchmark exports85%
775GLM 5.1OpenRouter harness, z-ai/glm-5.1-20260406, Crusoe, public 90-day aggregate· OpenRouter benchmark exports85%
776Qwen3.6 27BOpenRouter harness, qwen/qwen3.6-27b-20260422, Chutes, public 90-day aggregate· OpenRouter benchmark exports85%
777Step 3.5 FlashStep 3.5 technical report · Table 5 · PaCoRe· Reported by StepFun· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.85%
778GLM 5.3OpenRouter harness, z-ai/glm-5.3-20260816, BaseTen, public 90-day aggregate· OpenRouter benchmark exports85%
779GLM 5.3OpenRouter harness, z-ai/glm-5.3-20260816, River, public 90-day aggregate· OpenRouter benchmark exports85%
780Gemma 4 31B ITOpenRouter harness, google/gemma-4-31b-it-20260402, NextBit, public 90-day aggregate· OpenRouter benchmark exports85%
781DeepSeek V4 FlashOpenRouter harness, deepseek/deepseek-v4-flash-20260423, Venice, public 90-day aggregate· OpenRouter benchmark exports84.9%
782Nemotron 3.5 Super VLOpenRouter harness, nvidia/nemotron-3.5-super-vl-20261015, Venice, public 90-day aggregate· OpenRouter benchmark exports84.9%
783GLM 5.3OpenRouter harness, z-ai/glm-5.3-20260816, SambaNova, public 90-day aggregate· OpenRouter benchmark exports84.9%
784MiniMax M3.1 FlashOpenRouter harness, minimax/minimax-m3.1-flash-20261019, Fireworks, public 90-day aggregate· OpenRouter benchmark exports84.9%
785DeepSeek V4 Flash 0731OpenRouter harness, deepseek/deepseek-v4-flash-20260731, BaseTen, public 90-day aggregate· OpenRouter benchmark exports84.9%
786Gemma 4 31B ITOpenRouter harness, google/gemma-4-31b-it-20260402, ModelRun, public 90-day aggregate· OpenRouter benchmark exports84.9%
787MiMo V2.5OpenRouter harness, xiaomi/mimo-v2.5-20260422, Venice, public 90-day aggregate· OpenRouter benchmark exports84.9%
788Ling 3.0 Flash VLOpenRouter harness, inclusionai/ling-3.0-flash-vl-20260910, auto-routing, public 90-day aggregate· OpenRouter benchmark exports84.8%
789GPT 5.4low· Epoch AI Benchmarking Hub data84.8%
790Kimi K3low· Epoch AI Benchmarking Hub data84.8%
791Qwen3.5 Plus· Epoch AI Benchmarking Hub data84.8%
792Qwen3.6 27Bnone· Epoch AI Benchmarking Hub data84.8%
793Qwen3.6 35B A3Bnone· Epoch AI Benchmarking Hub data84.8%
794Qwen3.6 27BOpenRouter harness, qwen/qwen3.6-27b-20260422, Alibaba, public 90-day aggregate· OpenRouter benchmark exports84.8%
795GLM 5.1OpenRouter harness, z-ai/glm-5.1-20260406, Phala, public 90-day aggregate· OpenRouter benchmark exports84.8%
796GLM 5.3OpenRouter harness, z-ai/glm-5.3-20260816, ModelRun, public 90-day aggregate· OpenRouter benchmark exports84.8%
797GLM 5.3OpenRouter harness, z-ai/glm-5.3-20260816, AtlasCloud, public 90-day aggregate· OpenRouter benchmark exports84.8%
798Qwen3.6 27BOpenRouter harness, qwen/qwen3.6-27b-20260422, CoreWeave, public 90-day aggregate· OpenRouter benchmark exports84.8%
799Qwen3.6 35B A3BOpenRouter harness, qwen/qwen3.6-35b-a3b-20260415, AtlasCloud, public 90-day aggregate· OpenRouter benchmark exports84.8%
800Qwen3.6 27BOpenRouter harness, qwen/qwen3.6-27b-20260422, auto-routing, public 90-day aggregate· OpenRouter benchmark exports84.8%

Showing 751–800 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)