BenchAtlas

Rankings / Alibaba

Qwen3 32B

open weights

released 2025-04-28

BenchAtlas Index

as of 2026-09-11
28.9
thinking
rank #387
10 families · 5 categories · high
19.3
no reasoning
rank #438
10 families · 5 categories · high

Benchmark evidence

48 results
agentic coding
τ²-Benchthinking29.8 %independentArtificial Analysis
τ²-Bench
subset=banking
thinking5.4 %independentArtificial Analysis
coding
LiveCodeBench v6
implementation=artificial-analysis
thinking54.6 %independentArtificial Analysis
LiveCodeBench v6
implementation=artificial-analysis
no reasoning28.8 %independentArtificial Analysis
SciCodethinking36.0 %independentArtificial Analysis
SciCodeno reasoning28.0 %independentArtificial Analysis
external indices
AA Coding Indexthinking15.3 pointsindependentArtificial Analysis
Intelligence Index v4.1no reasoning7.3 pointsindependentArtificial Analysis
Intelligence Index v4.1thinking7.2 pointsindependentArtificial Analysis
AA Math Indexthinking73.0 pointsindependentArtificial Analysis
AA Math Indexno reasoning19.7 pointsindependentArtificial Analysis
ECI138.5 pointsindependentEpoch AI Benchmarking Hub
ECIno reasoning138.5 pointsindependentEpoch AI Benchmarking Hub
human preference
Coding (style control)older version
arena=text · category=coding · style_control=true
1406.9
1382.6–1431.2
communityLMArena Leaderboard Dataset
Industry Software And It Services (style control)older version
arena=text · category=industry_software_and_it_services · style_control=true
1398.0
1378.8–1417.3
communityLMArena Leaderboard Dataset
Hard Prompts English (style control)older version
arena=text · category=hard_prompts_english · style_control=true
1381.7
1360.8–1402.5
communityLMArena Leaderboard Dataset
Industry Life And Physical And Social Science (style control)older version
arena=text · category=industry_life_and_physical_and_social_science · style_control=true
1378.8
1358.4–1399.2
communityLMArena Leaderboard Dataset
Hard Prompts (style control)older version
arena=text · category=hard_prompts · style_control=true
1367.0
1350.9–1383.2
communityLMArena Leaderboard Dataset
English (style control)older version
arena=text · category=english · style_control=true
1365.4
1353.0–1377.8
communityLMArena Leaderboard Dataset
Longer Query (style control)older version
arena=text · category=longer_query · style_control=true
1354.9
1330.3–1379.5
communityLMArena Leaderboard Dataset
Overall (style control)
arena=text · category=overall · style_control=true
1346.9
1337.5–1356.4
communityLMArena Leaderboard Dataset
Multi Turn (style control)older version
arena=text · category=multi_turn · style_control=true
1337.9
1314.5–1361.2
communityLMArena Leaderboard Dataset
Instruction Following (style control)older version
arena=text · category=instruction_following · style_control=true
1331.1
1312.1–1350.1
communityLMArena Leaderboard Dataset
Non English (style control)older version
arena=text · category=non_english · style_control=true
1323.3
1310.0–1336.6
communityLMArena Leaderboard Dataset
Industry Writing And Literature And Language (style control)older version
arena=text · category=industry_writing_and_literature_and_language · style_control=true
1307.5
1290.5–1324.5
communityLMArena Leaderboard Dataset
Exclude Ties (style control)older version
arena=text · category=exclude_ties · style_control=true
1306.8
1292.2–1321.4
communityLMArena Leaderboard Dataset
Creative Writing (style control)older version
arena=text · category=creative_writing · style_control=true
1303.8
1281.2–1326.4
communityLMArena Leaderboard Dataset
Industry Entertainment And Sports And Media (style control)older version
arena=text · category=industry_entertainment_and_sports_and_media · style_control=true
1285.2
1264.6–1305.9
communityLMArena Leaderboard Dataset
knowledge science
GPQA Diamond
implementation=artificial-analysis
thinking66.8 %independentArtificial Analysis
GPQA Diamond65.7 %independentEpoch AI Benchmarking Hub
2026-08-28
GPQA Diamondno reasoning54.1 %independentEpoch AI Benchmarking Hub
2026-08-30
GPQA Diamond
implementation=artificial-analysis
no reasoning53.5 %independentArtificial Analysis
Humanity's Last Exam
implementation=artificial-analysis
thinking7.4 %independentArtificial Analysis
Humanity's Last Exam
implementation=artificial-analysis
no reasoning4.1 %independentArtificial Analysis
MMLU-Pro
implementation=artificial-analysis
thinking79.8 %independentArtificial Analysis
MMLU-Pro
implementation=artificial-analysis
no reasoning72.7 %independentArtificial Analysis
long context instruction
AA-LCRno reasoning0.0 %independentArtificial Analysis
AA-LCRthinking0.0 %independentArtificial Analysis
IFBenchthinking36.3 %independentArtificial Analysis
IFBenchno reasoning31.5 %independentArtificial Analysis
reasoning math
AIME
implementation=artificial-analysis
thinking80.7 %independentArtificial Analysis
AIME
year=2025 · implementation=artificial-analysis
thinking73.0 %independentArtificial Analysis
AIME
implementation=artificial-analysis
no reasoning30.3 %independentArtificial Analysis
AIME
year=2025 · implementation=artificial-analysis
no reasoning19.7 %independentArtificial Analysis
MATH-500
implementation=artificial-analysis
thinking96.1 %independentArtificial Analysis
MATH-500
implementation=artificial-analysis
no reasoning86.9 %independentArtificial Analysis
OTIS Mock AIME 2024–202566.9 %independentEpoch AI Benchmarking Hub
2026-08-30
OTIS Mock AIME 2024–2025no reasoning23.1 %independentEpoch AI Benchmarking Hub
2026-08-30

Agent + model results

systems, not bare-model scores
agent + model terminalagent + Qwen3 32BTerminal-Bench 1.015.5 %unverifiedTerminal-Bench Leaderboard
agent + model Artificial Analysis harness + Qwen3 32BTerminal-Bench 2.15.2 %independentArtificial Analysis
agent + model Artificial Analysis harness + Qwen3 32BTerminal-Bench Hard3.0 %independentArtificial Analysis

These scores measure the whole agent system (scaffold, tools, budgets) — they are never merged into the bare model’s numbers.

Related variants

Qwen3 7Qwen3 6 27BQwen3 5 27BQwen3 235B