Knowledge Reasoning on GPQA Diamond (Accuracy)

91.9Accuracy

Gemini 3 Pro

Updated 1mo ago

Evaluation Results

Method	Links
Gemini 3 Pro 2026.06		91.9
Qwen3.6 Plus 2026.06		90.4
Qwen3.5-397B-A17B 2026.06		88.4
Kimi K2.5 2026.06		87.6
Grok 4 2026.06		87.5
MiniMax M2.7 2026.06		87
Claude Opus 4.5 2026.06		87
Gemini 2.5 Pro 2026.06		86.4
GLM-5 2026.06		86
GPT-5 (high) 2026.06		85.7
MiMo v2 Flash 2026.06		83.7
OpenAI o3 (high) 2026.06		83.3
Gemini 2.5 Flash 2026.06		82.8
DeepSeek V3.2 2026.06		82.4
Phi4-Reasoning-Plus 2026.06		81.9
LongCat Flash 2026.06		81.5
Qwen3-235B-A22B-Thinking 2026.06		81.1
DeepSeek R1 0528 2026.06		81
GPT-OSS (high) 2026.06		80.1
Gemma-4-it 2026.06		78.8
Qwen3.5-4B 2026.06		76.2
GLM-4.5-Air 2026.06		75
Nemotron-3-Nano 2026.06		73
VibeThinker-3B + CLR 2026.06		72.9
VibeThinker-3B 2026.06		72.9
GPT-OSS-20B (high) 2026.06		71.5
Ministral-3-Reasoning-2512 2026.06		71.2
GPT-5 Nano (high) 2026.06		71.2
VibeThinker-3B 2026.06		70.2
VibeThinker-3B 2026.06		70.2
Qwen3-4B-Thinking-2507 2026.06		65.8
General Teacher 2026.05		62.5
Medical Teacher 2026.05		62.37
CaMOPD 2026.05		61.99
Relaxed OPD 2026.05		61.49
Hunyuan-4B-Instruct 2026.06		61.1
OpenReasoning-Nemotron 2026.06		61.1
Qwen2.5-32B-Instruct + Bootcamp-SFT-RL 2025.08		60.7
Mimo7B-RL-0530 2026.06		60.6
Vanilla MOPD 2026.05		60.1
SelecTKD 2026.05		58.46
DS-R1-Distilled-Qwen-32B + Bootcamp-RL 2025.08		51.6
Olmo-3-Think 2026.06		46.2
Qwen2.5-32B-Instruct 2025.08		44.7
Qwen2.5-32B-Instruct + Bootcamp-RL 2025.08		44.7
SmolLM3 2026.06		41.7
DS-R1-Distilled-Qwen-32B 2025.08		41.6
Qwen2.5-32B-Instruct + Bootcamp-SFT 2025.08		26.2