Qwen3 32B — Benchmarks

Benchmark scores for Qwen3 32B aggregated from public leaderboards (9 boards), with how it ranks among open models and how it compares with its nearest neighbours. See hardware requirements for what you need to run it.

llmrun Composite Score →

98.6

90% range 95.7–101.4 · from 8 boards

Coding93.190% range 87–99
AgentsNot enough data yet
MathNot enough data yet
Science96.090% range 89–103
Reasoning91.590% range 83–100

Qwen3 32B reasoning benchmarks

Multi-step logic and problem-solving — can the model think through a hard problem rather than recall an answer.

GPQA Diamond
GPQA Diamond is a set of extremely hard, graduate-level science questions (physics, chemistry, biology) written by domain experts and filtered so that skilled non-experts with web access still fail. It measures genuine reasoning rather than memorization.
65.7%
65.7%
median 54.0%open leader 93.1%
Overall best: Gemini 4 Argon (high) 96.4%
63rd percentile · #37 of 98 open · #190 of 300 all
#37 of 98
#190 of 300
Independent
Epoch AI · Apr 2025
CritPt
CritPt (Critical Points) tests graduate-level physics problem-solving on research-style questions spanning mechanics, electromagnetism, quantum mechanics, and more. It was created by physicists to probe whether a model can reason through multi-step physics rather than pattern-match textbook answers.
0.3%
0.3%
median 1.1%open leader 26.6%
Overall best: GPT 5.6 Sol Max 32.3%
36th percentile · #43 of 66 open · #177 of 219 all
#43 of 66
#177 of 219
Lab-reported
Epoch AI · Apr 2025
DTBench
DTBench is a set of handcrafted, expert-validated multiple-choice questions on decision theory — Newcomb-like problems and related puzzles — built by the Conceptual Reasoning Index project with Anthropic. It tests whether a model can follow an abstract argument rather than recall a fact.
67.5%
67.5%
median 67.7%open leader 93.9%
Overall best: Claude Opus 5.5 Max 98.9%
49th percentile · #35 of 68 open · #148 of 214 all
#35 of 68
#148 of 214
Lab-reported
Epoch AI · Apr 2025
Fiction.LiveBench
Fiction.LiveBench is a long-context comprehension test: models answer questions that require tracking characters, events and relationships across a long story. llmrun shows the 16k-token score, the column Epoch AI reports as its headline.
74.2%
74.2%
median 61.1%open leader 86.1%
Overall best: GPT 5 (Aug 07, 2025, medium) 97.2%
79th percentile · #6 of 25 open · #16 of 58 all
#6 of 25
#16 of 58
Independent
Epoch AI · Apr 2025
LMCA
LMCA (Language Model Conceptual Argumentation) scores models on expert-rated arguments across philosophy, decision theory, and AI risk — domains where there is no empirical feedback loop to check an answer against. It is part of the Conceptual Reasoning Index, built with Anthropic by researchers Emery Cooper and Caspar Oesterheld.
17.3%
17.3%
median 24.7%open leader 55.5%
Overall best: Claude Opus 5.5 Max 68.2%
32nd percentile · #37 of 54 open · #147 of 176 all
#37 of 54
#147 of 176
Lab-reported
Epoch AI · Apr 2025

Qwen3 32B coding benchmarks

Writing, editing, and fixing real code — measured by whether the resulting program actually runs and passes tests.

Aider Polyglot
The Aider Polyglot benchmark measures real-world coding across several programming languages: the model edits code to solve Exercism exercises, and is scored on whether the final solution actually runs and passes the tests.
40.0%
40.0%
median 45.1%open leader 74.2%
Overall best: GPT 5 (Aug 07, 2025, high) 88.0%
41st percentile · #11 of 18 open · #46 of 69 all
#11 of 18
#46 of 69
Independent
Epoch AI · Apr 2025
SciCode
SciCode asks a model to write real scientific-computing code — implementing multi-step numerical methods drawn from physics, chemistry, biology, and math research — and checks the output against reference solutions. It was built by a team of scientists and ML researchers to test coding ability on genuine research problems, not typical software tasks.
35.4%
35.4%
median 38.7%open leader 60.9%
Overall best: Claude Opus 5.5 Max 66.9%
31st percentile · #46 of 66 open · #182 of 207 all
#46 of 66
#182 of 207
Lab-reported
Epoch AI · Apr 2025

Qwen3 32B math benchmarks

Symbolic and quantitative problem-solving, from competition math to multi-step calculation.

AIME 2024/2025
AIME (American Invitational Mathematics Examination) is a prestigious high-school competition of hard, integer-answer problems. It's a widely-cited yardstick for multi-step mathematical reasoning, here on the 2024–2025 papers.
66.9%
66.9%
median 52.8%open leader 98.6%
Overall best: Claude Fable 5 (high) 100.0%
65th percentile · #31 of 86 open · #150 of 283 all
#31 of 86
#150 of 283
Independent
Epoch AI · Apr 2025

Qwen3 32B instruction following benchmarks

How precisely a model obeys explicit, verifiable instructions — format, length, and content constraints — and, more broadly, how much people prefer its responses head-to-head in blind human voting.

LMArena Text
Arena (arena.ai, formerly LMArena / Chatbot Arena) ranks models by millions of blind, head-to-head human votes on real prompts. The Arena Score shown here is a Bradley-Terry rating computed with Arena's default style control, so higher means people preferred that model's answers more often — it's a relative rating, not a percentage, so compare it only to other models on this board.
1346.8
1346.8
median 1321.5open leader 1487.1
Overall best: Gemini 4 Argon (High) 1523.9
58th percentile · #79 of 187 open · #214 of 405 all
#79 of 187
#214 of 405
Arena
Arena · Oct 2026

Score vs VRAM in the 24 GB tier

Qwen3 32B
Other models
Best Score for its VRAM
x: VRAM (GB) · y: llmrun Score

#11 of 35 that fit 24 GB

Each dot is a model at the best quant that fits a 24 GB GPU. The line joins the models with the best Score for their VRAM; a dot below it is beaten by a model that needs no more memory.

Full 24 GB ranking →

Qwen3 32B vs its neighbours

Same boards for the nearest open models per llmrun Score and the best model that fits 24 GB. The highest score in each row is marked.

Board	Qwen3 32Bthis model	Qwen3.8 27Bbest in classQ4_K_M · 18 GB	Qwen3 14Bnearest by ScoreQ4_K_M · 12 GB	DeepSeek R1 Distill Qwen 32Bnearest by ScoreQ4_K_S · 23 GB	Qwen3 30B A3B Instruct 2507nearest by ScoreQ4_K_M · 19 GB
llmrun Score	98.6	122.1	98.0	97.2	96.7
GPQA Diamond	65.7%	—	63.8%	64.1%	55.6%
CritPt	0.3%	5.4%	0.0%	—	—
DTBench	67.5%	88.0%	64.0%	—	67.2%
Fiction.LiveBench	74.2%	—	62.5%	—	—
LMCA	17.3%	41.4%	18.2%	—	22.4%
SciCode	35.4%	46.6%	31.6%	—	—
AIME 2024/2025	66.9%	—	66.4%	55.6%	62.2%
LMArena Text	1346.8	1437.8	—	—	1382.3
Compare head to head		Compare	Compare	Compare	Compare

Scores aggregated from public benchmark sources (Epoch AI: 8 boards, Arena: 1 board) — each linked from the benchmark pages. llmrun does not run these benchmarks. Percentiles and ranks are among open models in the llmrun catalogue; the median and leader marks refer to the same field.