Hallusion Bench Benchmark

Hallusion Bench Benchmark | Phaseo

Models Using This Benchmark

Organisation	Model	Reported	Top Score	Info	Self Reported	Source
Qwen	Qwen 3.5 27B	24 Feb 2026	70%	-	Yes	Source
Qwen	Qwen 3.5 Flash	23 Feb 2026	70%	inferred family alias from qwen3.5-27b (score=0.4147; benches=81)	Yes	Source
Qwen	Qwen 3.5 35B A3B	24 Feb 2026	67.90%	-	Yes	Source
Qwen	Qwen 3.5 122B A10B	24 Feb 2026	67.60%	-	Yes	Source
Qwen	Qwen 3 VL 32B Thinking	-	67.40%	-	Yes	-
Qwen	Qwen 3 VL 235B A22B Thinking	-	66.70%	-	Yes	-
Qwen	Qwen 3 VL 30B A3B Thinking	-	66%	-	Yes	-
Qwen	Qwen 3 8B	-	65.40%	inferred high-confidence family alias from qwen3-vl-8b-thinking (score=0.4600; benches=50)	Yes	-
Qwen	Qwen 3 Embedding 8B	-	65.40%	inferred family alias from qwen3-vl-8b-thinking (score=0.3850; benches=50)	Yes	-
Qwen	Qwen 3 Guard Gen 8B	-	65.40%	inferred family alias from qwen3-vl-8b-thinking (score=0.3400; benches=50)	Yes	-
Qwen	Qwen 3 Guard Stream 8B	-	65.40%	inferred family alias from qwen3-vl-8b-thinking (score=0.3371; benches=50)	Yes	-
Qwen	Qwen 3 Reranker 8B	-	65.40%	inferred family alias from qwen3-vl-8b-thinking (score=0.3850; benches=50)	Yes	-
Qwen	Qwen 3 VL 8B Thinking	-	65.40%	-	Yes	-
Qwen	Qwen 3 VL Embedding 8B	-	65.40%	inferred high-confidence family alias from qwen3-vl-8b-thinking (score=0.5232; benches=50)	Yes	-
Qwen	Qwen 3 VL Reranker 8B	-	65.40%	inferred high-confidence family alias from qwen3-vl-8b-thinking (score=0.5275; benches=50)	Yes	-
Qwen	Qwen 3 4B Instruct 2507	-	64.10%	inferred family alias from qwen3-vl-4b-thinking (score=0.3462; benches=48)	Yes	-
Qwen	Qwen 3 4B SafeRL	-	64.10%	inferred family alias from qwen3-vl-4b-thinking (score=0.3850; benches=48)	Yes	-
Qwen	Qwen 3 4B Thinking 2507	-	64.10%	inferred family alias from qwen3-vl-4b-thinking (score=0.3462; benches=48)	Yes	-
Qwen	Qwen 3 Embedding 4B	-	64.10%	inferred family alias from qwen3-vl-4b-thinking (score=0.3850; benches=48)	Yes	-
Qwen	Qwen 3 Guard Gen 4B	-	64.10%	inferred family alias from qwen3-vl-4b-thinking (score=0.3400; benches=48)	Yes	-
Qwen	Qwen 3 Guard Stream 4B	-	64.10%	inferred family alias from qwen3-vl-4b-thinking (score=0.3371; benches=48)	Yes	-
Qwen	Qwen 3 Reranker 4B	-	64.10%	inferred family alias from qwen3-vl-4b-thinking (score=0.3850; benches=48)	Yes	-
Qwen	Qwen 3 VL 4B Thinking	-	64.10%	-	Yes	-
Qwen	Qwen 3 VL 32B Instruct	-	63.80%	-	Yes	-
Qwen	Qwen 3 VL 235B A22B Instruct	-	63.20%	-	Yes	-
Qwen	Qwen 3 VL 30B A3B Instruct	-	61.50%	-	Yes	-
Qwen	Qwen 3 VL 8B Instruct	-	61.10%	-	Yes	-
Qwen	Qwen 3 VL 4B Instruct	-	57.60%	-	Yes	-
Qwen	Qwen 2.5 Math 72B	-	55.16%	inferred high-confidence family alias from qwen2.5-vl-72b (score=0.4700; benches=30)	Yes	Source
Qwen	Qwen 2.5 Math PRM 72B	-	55.16%	inferred family alias from qwen2.5-vl-72b (score=0.4050; benches=30)	Yes	Source
Qwen	Qwen 2.5 Math RM 72B	-	55.16%	inferred family alias from qwen2.5-vl-72b (score=0.4092; benches=30)	Yes	Source
Qwen	Qwen 2.5 VL 72B Instruct	26 Jan 2025	55.16%	-	Yes	Source
Qwen	Qwen 2.5 0.5B	-	52.90%	inferred family alias from qwen2.5-vl-7b (score=0.3131; benches=32)	Yes	Source
Qwen	Qwen 2.5 1.5B	-	52.90%	inferred family alias from qwen2.5-vl-7b (score=0.3131; benches=32)	Yes	Source
Qwen	Qwen 2.5 3B	-	52.90%	inferred family alias from qwen2.5-vl-7b (score=0.3531; benches=32)	Yes	Source
Qwen	Qwen 2.5 VL 7B Instruct	-	52.90%	-	Yes	Source

Average Score

Score Range

Leading Model

Hallusion Bench

Recorded Results

Average Score

Score Range

Leading Model

Models Using This Benchmark