Qwen3-32B

Qwen3-32B ranks first on ALRAGE.

Qwen Released 28 Apr 2025

Benchmarks with a score

Ranked first, out of five or more models

In the top three

Qwen Apr 2025

Open weights

138.5 #133 of 271

Epoch Capabilities Index

Where it ranks

Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.

Benchmark	Score	Rank	Source
ALRAGE	80.7%	#1 of 39	Official leaderboard
Arena-Hard-Auto	93.8%	#4 of 39	Model card
AlignBench	8.72	#3 of 15	Model card
Vectara Hallucination Leaderboard (2025 dataset)	5.90%	#24 of 104	Official leaderboard
AgentIF	58.4%	#3 of 12	Official leaderboard
MathIF	43.8%	#3 of 12	Official leaderboard
MLogiQA	76.3%	#2 of 8	Model card
Fiction.LiveBench	74.2%	#16 of 57	Epoch AI
MT-AIME2024	75%	#2 of 7	Model card
RULER	93.7%	#11 of 38	Model card
Creative Writing v3	81.0	#5 of 12	Model card
Multi-IF	73%	#13 of 28	Model card
LongBench v2	49.2%	#16 of 33	Official leaderboard
MMLU-Redux	90.9%	#20 of 38	Model card
LMArena (Elo)	1,347	#179 of 333	Official leaderboard
SEA-HELM	67.7	#35 of 62	Official leaderboard
Arena-Hard v2.0	44.5%	#13 of 23	Official leaderboard
WritingBench	7.90	#8 of 14	Model card
C-Eval	87.3%	#22 of 38	Model card
IFEval	85%	#73 of 126	Model card
INCLUDE	73.7%	#15 of 25	Model card
MMMLU	80.6%	#39 of 62	Model card
AlGhafa	46.8%	#28 of 39	Official leaderboard
ArbMMLU-HT	29.5%	#29 of 39	Official leaderboard
MadinahQA	33%	#29 of 39	Official leaderboard
Open Arabic LLM Leaderboard v2	41.2%	#29 of 39	Official leaderboard
Vectara Hallucination Leaderboard (HHEM)	2.80%	#113 of 148	Official leaderboard
EQ-Bench 3	1,069	#60 of 78	Official leaderboard
ArabicMMLU	31.5%	#32 of 39	Official leaderboard
Longform Creative Writing (EQ-Bench)	40.5	#99 of 119	Official leaderboard
CounterFactQA (Qwen)	68.1%	#1 of 1	Model card
LengthCtrl (Qwen)	73.5%	#1 of 1	Model card
ThinkFollow (Qwen)	98.9%	#1 of 1	Model card

Benchmark	Score	Rank	Source
MATH-500	97.2%	#18 of 67	Model card
ZebraLogic	88.8%	#9 of 23	Model card
AIME 2024	81.4%	#36 of 87	Model card
HiPhO	55.4%	#17 of 35	Official leaderboard
OTIS Mock AIME 2024-2025	66.9%	#106 of 202	Epoch AI
GPQA Diamond	65.7%	#117 of 219	Epoch AI
PolyMath	47.4%	#12 of 18	Model card
CritPt	0.29%	#107 of 160	Epoch AI
DTBench	45.8%	#120 of 178	Official leaderboard
AIME 2025	72.9%	#104 of 143	Model card
AutoLogi	87.3%	#7 of 8	Model card

Benchmark	Score	Rank	Source
LiveCodeBench v5	65.7%	#16 of 44	Model card
AutoCodeBench	41.7%	#20 of 34	Official leaderboard
Aider Polyglot	40%	#33 of 54	Epoch AI
Codeforces rating	1,977	#26 of 41	Model card

Benchmark	Score	Rank	Source
BFCL V4	48.7%	#22 of 65	Official leaderboard
Berkeley Function Calling Leaderboard (BFCL)	70.3%	#23 of 47	Model card
SnakeBench	24	#91 of 183	Official leaderboard
GAIA	21.6%	#47 of 57	Official leaderboard
LisanBench	68.0	#119 of 133	Official leaderboard
VITA-Bench	5.30%	#18 of 18	Official leaderboard
ToolUse (Qwen)	85.5%	#1 of 1	Model card

Benchmark	Score		Rank	Source
LEXam	40%		#23 of 31	Official leaderboard

Benchmark	Score		Rank	Source
AraTrust	38.9%		#34 of 39	Official leaderboard

Benchmark	Score	Rank	Source
LiveBench	74.9%	#8 of 27	Model card
Political Science LLM Benchmark	64.8%	#20 of 29	Official leaderboard
ACCoRD	48.2	#115 of 150	Official leaderboard
Conceptual Reasoning Index	29.2	#115 of 149	Official leaderboard
LMCA	17.3%	#111 of 139	Epoch AI

Qwen's models in color, against every other model on the Epoch Capabilities Index, by release date.

Every result in the longer documents the lab published with it. 29 of them aren't in the launch post.