Qwen2-72B

Released2024-06-07
Price per 1MNot listed
Overall score31.2 · rank 230
ProviderAlibaba
WeightsOpen weights

Scores

DimensionScoreRankBenchmarks
Overall: Balanced 31.2 230 4
Overall: Coding 30.8 227 2
Overall: Reasoning 33.7 224 2
Overall: Agentic work 30.2 231 2
Agents 29.3 185 1
Reasoning 33.9 224 1
Math 33.4 187 1
General 28.9 183 1

Evidence

Agents 29.3 ±3.3

BenchmarkResultSourceDate
METR task suite 29.9% Epoch AI 2026-11-11

Reasoning 33.9 ±4.6

BenchmarkResultSourceDate
GPQA Diamond 40.8% Epoch AI 2025-01-27

Math 33.4 ±2.4

BenchmarkResultSourceDate
MATH Level 5 39.1% Epoch AI 2025-01-27

General 28.9 ±4.1

BenchmarkResultSourceDate
LMArena Text 1,262 LMArena 2026-11-09

Names matched to this model

qwen2-72b, qwen2-72b-instruct

Compare Qwen2-72B with other models