SuperGPQA

SuperGPQA - Benchmark Leaderboard & Model Performance | AI Stats

Models Using This Benchmark

Organisation	Model	Reported	Top Score	Info	Self Reported	Source
Qwen	Qwen 3.6 Plus	01 Apr 2026	71.60%	-	Yes	Source
Qwen	Qwen 3.5 397B A17B	16 Feb 2026	70.40%	-	Yes	Source
ByteDance	Seed 2.0 Pro	14 Feb 2026	68.70%	Seed2 official benchmark table \| SuperGPQA	Yes	Source
ByteDance	Seed 2.0 Lite	14 Feb 2026	67.50%	Seed2 official benchmark table \| SuperGPQA	Yes	Source
Qwen	Qwen 3.5 122B A10B	24 Feb 2026	67.10%	-	Yes	Source
Qwen	Qwen 3.5 27B	24 Feb 2026	65.60%	-	Yes	Source
Qwen	Qwen 3.5 Flash	23 Feb 2026	65.60%	inferred family alias from qwen3.5-27b (score=0.4147; benches=81)	Yes	Source
Qwen	Qwen 3 TTS (2025-11-27)	-	65.10%	inferred family alias from qwen3-max (score=0.3833; benches=6)	Yes	Source
Qwen	Qwen 3 14B	-	65.10%	inferred family alias from qwen3-max (score=0.3333; benches=6)	Yes	Source
Qwen	Qwen 3 Max Thinking	26 Jan 2026	65.10%	inferred alias from qwen3-max	Yes	Source
Qwen	Qwen 3 235B A22B Thinking 2507	-	64.90%	-	Yes	Source
Qwen	Qwen 3 VL 235B A22B Thinking	-	64.30%	-	Yes	-
Qwen	Qwen 3.5 35B A3B	24 Feb 2026	63.40%	-	Yes	Source
Qwen	Qwen 3 A235 A22B Instruct 2507	-	62.60%	-	Yes	Source
ByteDance	Seed 2.0 Mini	14 Feb 2026	61.60%	Seed2 official benchmark table \| SuperGPQA	Yes	Source
Qwen	Qwen 3 Next 80B A3B Thinking	-	60.80%	-	Yes	Source
Qwen	Qwen 3 VL 235B A22B Instruct	-	60.40%	-	Yes	-
Qwen	Qwen 3 VL 32B Thinking	-	59%	-	Yes	-
Qwen	Qwen 3 Next 80B A3B Instruct	-	58.80%	-	Yes	Source
Qwen	Qwen 3.5 9B	02 Mar 2026	58.20%	-	Yes	Source
Moonshot	Kimi K2 (2025-09-05)	05 Sept 2025	57.20%	Acc	Yes	Source
Qwen	Qwen 3 VL 30B A3B Thinking	-	56.40%	-	Yes	-
Qwen	Qwen 3 VL 32B Instruct	-	54.60%	-	Yes	-
Qwen	Qwen 3 VL 30B A3B Instruct	-	53.10%	-	Yes	-
Qwen	Qwen 3.5 4B	02 Mar 2026	52.90%	-	Yes	Source
Qwen	Qwen 3 VL Embedding 8B	-	51.20%	inferred high-confidence family alias from qwen3-vl-8b-thinking (score=0.5232; benches=50)	Yes	-
Qwen	Qwen 3 Reranker 8B	-	51.20%	inferred family alias from qwen3-vl-8b-thinking (score=0.3850; benches=50)	Yes	-
Qwen	Qwen 3 Guard Gen 8B	-	51.20%	inferred family alias from qwen3-vl-8b-thinking (score=0.3400; benches=50)	Yes	-
Qwen	Qwen 3 Embedding 8B	-	51.20%	inferred family alias from qwen3-vl-8b-thinking (score=0.3850; benches=50)	Yes	-
Qwen	Qwen 3 8B	-	51.20%	inferred high-confidence family alias from qwen3-vl-8b-thinking (score=0.4600; benches=50)	Yes	-
Qwen	Qwen 3 Guard Stream 8B	-	51.20%	inferred family alias from qwen3-vl-8b-thinking (score=0.3371; benches=50)	Yes	-
Qwen	Qwen 3 VL 8B Thinking	-	51.20%	-	Yes	-
Qwen	Qwen 3 VL Reranker 8B	-	51.20%	inferred high-confidence family alias from qwen3-vl-8b-thinking (score=0.5275; benches=50)	Yes	-
Qwen	Qwen 3 Reranker 4B	-	46.80%	inferred family alias from qwen3-vl-4b-thinking (score=0.3850; benches=48)	Yes	-
Qwen	Qwen 3 Embedding 4B	-	46.80%	inferred family alias from qwen3-vl-4b-thinking (score=0.3850; benches=48)	Yes	-
Qwen	Qwen 3 VL 4B Thinking	-	46.80%	-	Yes	-
Qwen	Qwen 3 Guard Gen 4B	-	46.80%	inferred family alias from qwen3-vl-4b-thinking (score=0.3400; benches=48)	Yes	-
Qwen	Qwen 3 4B Instruct 2507	-	46.80%	inferred family alias from qwen3-vl-4b-thinking (score=0.3462; benches=48)	Yes	-
Qwen	Qwen 3 4B SafeRL	-	46.80%	inferred family alias from qwen3-vl-4b-thinking (score=0.3850; benches=48)	Yes	-
Qwen	Qwen 3 4B Thinking 2507	-	46.80%	inferred family alias from qwen3-vl-4b-thinking (score=0.3462; benches=48)	Yes	-
Qwen	Qwen 3 Guard Stream 4B	-	46.80%	inferred family alias from qwen3-vl-4b-thinking (score=0.3371; benches=48)	Yes	-
Moonshot	Kimi K2 (2025-07-11)	11 Jul 2025	44.70%	EM	Yes	Source
Qwen	Qwen 3 VL 8B Instruct	-	44.50%	-	Yes	-
Qwen	Qwen 3 235B A22B	-	44.06%	-	Yes	Source
Qwen	Qwen 3 VL 4B Instruct	-	40.30%	-	Yes	-
Qwen	Qwen 3.5 2B	02 Mar 2026	37.50%	-	Yes	Source
Qwen	Qwen 3.5 0.8B	02 Mar 2026	21.30%	-	Yes	Source

Recorded Results

Average Score

Score Range

Leading Model

Models Using This Benchmark