Qwen 3 Max

Model

Compare Models

Qwen 3 Max

Release Date: Sep 23, 2025

Developer Alibaba 🇨🇳

Context Window 262k

Max Output Tokens 66k

Token Costs (in/out) $1.20/6.00

Weights Private

Input Modalities

Accuracy

55.17 %

Avg. Cost (In/Out)

$ 1.20 / $ 6.00

Latency

13 min 49 s

BenchmarksAccuracyRankings

TaxEval v2

0.0%

±0.86

45/141

Vibe Code Bench v1.1

0.0%

±1.29

77/85

GPQA Diamond

0.0%

±2.03

67/133

IOI

0.0%

±6.35

28/62

LiveCodeBench

0.0%

±1.11

72/138

LegalBench

0.0%

±0.40

68/137

MMLU Pro

0.0%

±0.36

60/133

Proprietary BenchmarksAcademic BenchmarksIndustry Partners

View Hyperparameter Settings

Default Provider : Alibaba

Temperature: 0.7

Top P: Default

Top K: Default

Max Output Tokens: 65,536

Updates

Sep 25, 2025

Qwen 3 Max Evaluated

We evaluated Qwen 3 Max and found the following:

Qwen 3 Max breaks the top 20 on 5 benchmarks, with leading open source scores on LiveCodeBench, MMLU Pro, and GPQA Diamond.
Overall scores are similar to Qwen 3 Max Preview, with gains on just three benchmarks: FAB (+15%), IOI (+8%), and AIME (+17.2%). However, Qwen 3 Max Preview far outperformed Max on LegalBench (78.9% vs. 38.1%).
Our results were lower than Alibaba’s reported AIME25 score by 11.6%, though LiveCodeBench results aligned closely.

Overall, Qwen 3 Max is a solid, middle-of-the-pack model. It shows incremental gains over Qwen 3 Max Preview on FAB, IOI, and AIME, while underperforming on LegalBench.