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
101Claude Opus 5· Epoch AI Benchmarking Hub data92.9%
102Kimi K3.1 ProOpenRouter harness, moonshotai/kimi-k3.1-pro-20261023, AkashML, public 90-day aggregate· OpenRouter benchmark exports92.9%
103Claude Sonnet 5.5OpenRouter harness, anthropic/claude-sonnet-5.5-20260928, Anthropic, public 90-day aggregate· OpenRouter benchmark exports92.9%
104GPT-5.6 TerraOpenAI GPT-5.6 launch table· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.92.9%
105Kimi K3DeepSeek V4.1 launch chart · exact public value · setting not further specified· Reported by DeepSeek· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.92.9%
106Kimi K3Pass@1 / max effort· Reported by DeepSeek· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.92.9%
107GPT-5.6 SolOpenRouter harness, openai/gpt-5.6-sol-20260709, Azure, public 90-day aggregate· OpenRouter benchmark exports92.9%
108GPT 5.5OpenRouter harness, openai/gpt-5.5-20260423, Amazon Bedrock, public 90-day aggregate· OpenRouter benchmark exports92.8%
109Grok 4.6OpenRouter harness, x-ai/grok-4.6-20260810, auto-routing, public 90-day aggregate· OpenRouter benchmark exports92.8%
110Claude Sonnet 5.5OpenRouter harness, anthropic/claude-sonnet-5.5-20260928, auto-routing, public 90-day aggregate· OpenRouter benchmark exports92.8%
111DeepSeek V4.1 Flash 1017OpenRouter harness, deepseek/deepseek-v4.1-flash-20261017, Tenstorrent, public 90-day aggregate· OpenRouter benchmark exports92.8%
112Gemini 3.5 Flashhigh· Epoch AI Benchmarking Hub data92.8%
113GPT 5.4OpenAI GPT-5.4 launch table· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.92.8%
114Gemini 3.6 FlashOpenRouter harness, google/gemini-3.6-flash-20260721, auto-routing, public 90-day aggregate· OpenRouter benchmark exports92.8%
115Kimi K3.1 ProOpenRouter harness, moonshotai/kimi-k3.1-pro-20261023, Fireworks, public 90-day aggregate· OpenRouter benchmark exports92.8%
116Kimi K3OpenRouter harness, moonshotai/kimi-k3-20260715, Parasail, public 90-day aggregate· OpenRouter benchmark exports92.8%
117Grok 4.7xhigh· Epoch AI Benchmarking Hub data92.7%
118Kimi K3.1 ProOpenRouter harness, moonshotai/kimi-k3.1-pro-20261023, PrimaLabs, public 90-day aggregate· OpenRouter benchmark exports92.7%
119Kimi K3OpenRouter harness, moonshotai/kimi-k3-20260715, Nebius, public 90-day aggregate· OpenRouter benchmark exports92.7%
120Kimi K3.1 ProOpenRouter harness, moonshotai/kimi-k3.1-pro-20261023, Phala, public 90-day aggregate· OpenRouter benchmark exports92.7%
121GPT-5.6 SolOpenRouter harness, openai/gpt-5.6-sol-20260709, OpenAI, public 90-day aggregate· OpenRouter benchmark exports92.6%
122Gemini 3 Pro Preview· Epoch AI Benchmarking Hub data92.6%
123Claude Sonnet 5.5OpenRouter harness, anthropic/claude-sonnet-5.5-20260928, Amazon Bedrock, public 90-day aggregate· OpenRouter benchmark exports92.6%
124Kimi K3OpenRouter harness, moonshotai/kimi-k3-20260715, Morph, public 90-day aggregate· OpenRouter benchmark exports92.6%
125Claude Fable 5OpenAI GPT-5.6 launch table· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.92.6%
126Claude Fable 5OpenAI launch table· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.92.6%
127Claude Fable 5Qwen3.8 官方模型卡对照表· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.92.6%
128GPT-5.3 CodexOpenAI GPT-5.4 launch table· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.92.6%
129Qwen3.8 MaxQwen3.8 官方模型卡对照表· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.92.6%
130Kimi K3OpenRouter harness, moonshotai/kimi-k3-20260715, Fireworks, public 90-day aggregate· OpenRouter benchmark exports92.6%
131Kimi K3OpenRouter harness, moonshotai/kimi-k3-20260715, Decart, public 90-day aggregate· OpenRouter benchmark exports92.6%
132Kimi K3OpenRouter harness, moonshotai/kimi-k3-20260715, Relace, public 90-day aggregate· OpenRouter benchmark exports92.6%
133Kimi K3OpenRouter harness, moonshotai/kimi-k3-20260715, River, public 90-day aggregate· OpenRouter benchmark exports92.6%
134Gemini 3.5 FlashOpenRouter harness, google/gemini-3.5-flash-20260519, Google, public 90-day aggregate· OpenRouter benchmark exports92.6%
135GLM 5.3 FlashOpenRouter harness, z-ai/glm-5.3-flash-20260826, Darkbloom, public 90-day aggregate· OpenRouter benchmark exports92.5%
136Grok 4.6OpenRouter harness, x-ai/grok-4.6-20260810, xAI, public 90-day aggregate· OpenRouter benchmark exports92.5%
137GPT-6 SolOpenRouter harness, openai/gpt-6-sol-20260922, Azure, public 90-day aggregate· OpenRouter benchmark exports92.5%
138Kimi K2.7 CodeOpenRouter harness, moonshotai/kimi-k2.7-code-20260612, Venice, public 90-day aggregate· OpenRouter benchmark exports92.4%
139Kimi K3OpenRouter harness, moonshotai/kimi-k3-20260715, Wafer, public 90-day aggregate· OpenRouter benchmark exports92.4%
140Kimi K3OpenRouter harness, moonshotai/kimi-k3-20260715, Makora, public 90-day aggregate· OpenRouter benchmark exports92.4%
141GPT-5.6 SolOpenRouter harness, openai/gpt-5.6-sol-20260709, auto-routing, public 90-day aggregate· OpenRouter benchmark exports92.4%
142Kimi K3.1 ProOpenRouter harness, moonshotai/kimi-k3.1-pro-20261023, Together, public 90-day aggregate· OpenRouter benchmark exports92.4%
143DeepSeek V4 Pro 0813DeepSeek V4.1 launch chart · exact public value · setting not further specified· Reported by DeepSeek· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.92.4%
144DeepSeek V4 Pro 0813Pass@1 / max effort· Reported by DeepSeek· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.92.4%
145GPT 5.2no tools · xhigh· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.92.4%
146GPT 5.2OpenAI GPT-5.4 launch table· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.92.4%
147Qwen3.7 MaxQwen3.7-Max release · recommended reasoning-effort=xhigh system prompt· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.92.4%
148Qwen3.7 MaxQwen3.8 官方模型卡对照表· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.92.4%
149Kimi K3OpenRouter harness, moonshotai/kimi-k3-20260715, InferenceNet, public 90-day aggregate· OpenRouter benchmark exports92.4%
150DeepSeek V4.1 Flash 1017OpenRouter harness, deepseek/deepseek-v4.1-flash-20261017, Near AI, public 90-day aggregate· OpenRouter benchmark exports92.3%

Showing 101–150 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)