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
1151GPT 5.5none· Epoch AI Benchmarking Hub data77.3%
1152GPT-5.6 Terranone· Epoch AI Benchmarking Hub data77.3%
1153GPT OSS 120BOpenRouter harness, openai/gpt-oss-120b, SambaNova, public 90-day aggregate· OpenRouter benchmark exports77.2%
1154Gemma 4 26B A4BOpenRouter harness, google/gemma-4-26b-a4b-it-20260403, CoreWeave, public 90-day aggregate· OpenRouter benchmark exports77.1%
1155Gemma 4 26B A4BOpenRouter harness, google/gemma-4-26b-a4b-it-20260403, Novita, public 90-day aggregate· OpenRouter benchmark exports77.1%
1156MiMo V2.6 FlashOpenRouter harness, xiaomi/mimo-v2.6-flash-20260921, Xiaomi, public 90-day aggregate· OpenRouter benchmark exports77.1%
1157GPT OSS 120BOpenRouter harness, openai/gpt-oss-120b, AkashML, public 90-day aggregate· OpenRouter benchmark exports77.1%
1158GPT OSS 120BOpenRouter harness, openai/gpt-oss-120b, BaseTen, public 90-day aggregate· OpenRouter benchmark exports77%
1159Ling 3.0 FlashOpenRouter harness, inclusionai/ling-3.0-flash-20260723, Novita, public 90-day aggregate· OpenRouter benchmark exports77%
1160o3 Minihigh· Epoch AI Benchmarking Hub data77%
1161Gemma 4 26B A4BOpenRouter harness, google/gemma-4-26b-a4b-it-20260403, Parasail, public 90-day aggregate· OpenRouter benchmark exports77%
1162MiMo V2.5OpenRouter harness, xiaomi/mimo-v2.5-20260422, Parasail, public 90-day aggregate· OpenRouter benchmark exports77%
1163DeepSeek V3.1 TerminusOpenRouter harness, deepseek/deepseek-v3.1-terminus, auto-routing, public 90-day aggregate· OpenRouter benchmark exports76.9%
1164GPT OSS 120BOpenRouter harness, openai/gpt-oss-120b, Mara, public 90-day aggregate· OpenRouter benchmark exports76.8%
1165GPT OSS 120BOpenRouter harness, openai/gpt-oss-120b, Venice, public 90-day aggregate· OpenRouter benchmark exports76.8%
1166Gemma 4 26B A4BOpenRouter harness, google/gemma-4-26b-a4b-it-20260403, SiliconFlow, public 90-day aggregate· OpenRouter benchmark exports76.8%
1167Claude Opus 4.127k· Epoch AI Benchmarking Hub data76.8%
1168Claude Sonnet 3.716k· Epoch AI Benchmarking Hub data76.8%
1169Claude Sonnet 3.732k· Epoch AI Benchmarking Hub data76.8%
1170o1high· Epoch AI Benchmarking Hub data76.8%
1171GPT OSS 120BOpenRouter harness, openai/gpt-oss-120b, Nebius, public 90-day aggregate· OpenRouter benchmark exports76.8%
1172GPT OSS 120BOpenRouter harness, openai/gpt-oss-120b, Near AI, public 90-day aggregate· OpenRouter benchmark exports76.8%
1173GPT OSS 120BOpenRouter harness, openai/gpt-oss-120b, Amazon Bedrock, public 90-day aggregate· OpenRouter benchmark exports76.7%
1174Gemma 4 26B A4BOpenRouter harness, google/gemma-4-26b-a4b-it-20260403, DekaLLM, public 90-day aggregate· OpenRouter benchmark exports76.7%
1175Claude Sonnet 4OpenRouter harness, anthropic/claude-4-sonnet-20250522, Amazon Bedrock, public 90-day aggregate· OpenRouter benchmark exports76.7%
1176Gemini 2.5 FlashOpenRouter harness, google/gemini-2.5-flash, Google AI Studio, public 90-day aggregate· OpenRouter benchmark exports76.7%
1177GPT OSS 120BOpenRouter harness, openai/gpt-oss-120b, Crusoe, public 90-day aggregate· OpenRouter benchmark exports76.5%
1178MiMo V2.6 FlashOpenRouter harness, xiaomi/mimo-v2.6-flash-20260921, Tenstorrent, public 90-day aggregate· OpenRouter benchmark exports76.5%
1179GPT OSS 120BOpenRouter harness, openai/gpt-oss-120b, Sail Research, public 90-day aggregate· OpenRouter benchmark exports76.5%
1180Qwen 3.8 27BOpenRouter harness, qwen/qwen3.8-27b-20260814, Reka, public 90-day aggregate· OpenRouter benchmark exports76.5%
1181GPT OSS 120BOpenRouter harness, openai/gpt-oss-120b, Groq, public 90-day aggregate· OpenRouter benchmark exports76.5%
1182Qwen3.6 35B A3BOpenRouter harness, qwen/qwen3.6-35b-a3b-20260415, DeepInfra, public 90-day aggregate· OpenRouter benchmark exports76.5%
1183DeepSeek R1OpenRouter harness, deepseek/deepseek-r1-0528, auto-routing, public 90-day aggregate· OpenRouter benchmark exports76.4%
1184DeepSeek R1· Epoch AI Benchmarking Hub data
Values from 2 sources
76.3%
1185Trinity Large ThinkingTrinity-Large-Thinking Model Card· Reported by Arcee AI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.76.3%
1186Claude Sonnet 4OpenRouter harness, anthropic/claude-4-sonnet-20250522, Google, public 90-day aggregate· OpenRouter benchmark exports76.3%
1187GLM 5.1OpenRouter harness, z-ai/glm-5.1-20260406, GMICloud, public 90-day aggregate· OpenRouter benchmark exports76.3%
1188GPT OSS 120BOpenRouter harness, openai/gpt-oss-120b, Together, public 90-day aggregate· OpenRouter benchmark exports76.3%
1189Claude Opus 416k· Epoch AI Benchmarking Hub data76.3%
1190Grok 3 Minilow, beta· Epoch AI Benchmarking Hub data76.3%
1191Qwen3 Coder NextOpenRouter harness, qwen/qwen3-coder-next-2025-02-03, PrimeIntellect, public 90-day aggregate· OpenRouter benchmark exports76.3%
1192MiMo V2.6 FlashOpenRouter harness, xiaomi/mimo-v2.6-flash-20260921, InferenceNet, public 90-day aggregate· OpenRouter benchmark exports76.3%
1193Qwen3.5 4B· OpenEvals Leaderboard data· Reported by the model's developer. The source relaying it did not verify it.76.2%
1194Ling 3.0 FlashOpenRouter harness, inclusionai/ling-3.0-flash-20260723, DeepInfra, public 90-day aggregate· OpenRouter benchmark exports76.1%
1195GLM 5OpenRouter harness, z-ai/glm-5-20260211, Baidu, public 90-day aggregate· OpenRouter benchmark exports75.9%
1196Mistral SmallOpenRouter harness, mistralai/mistral-small-2603, auto-routing, public 90-day aggregate· OpenRouter benchmark exports75.8%
1197Claude Sonnet 416k· Epoch AI Benchmarking Hub data75.8%
1198Gemini 3.5 Flash Litelow· Epoch AI Benchmarking Hub data75.8%
1199Gemma 4 31B ITminimal· Epoch AI Benchmarking Hub data75.8%
1200GPT OSS 120Bhigh· Epoch AI Benchmarking Hub data75.8%

Showing 1,151–1,200 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)