Qwen3.7 Max

Name: Qwen3.7 Max
Author: Alibaba

Alibaba

Qwen3.7 Max is a proprietary flagship model from the Qwen team, released in May 2026. It leads Qwen's internal front-end generation benchmarks and tops HMMT Feb 26 with a 0.97 score, with strong agentic tool-use results on Kernel Bench L3. It is available via Novita and Together.

Key Specifications

Parameters

Context

1.0M

Release Date

May 19, 2026

Average Score

73.6%

API Documentation Research Paper Results Blog

Timeline

Key dates in the model's history

Announcement

May 19, 2026

Last Update

August 27, 2026

Today

September 10, 2026

Technical Specifications

Parameters

Training Tokens

Knowledge Cutoff

Family

Capabilities

MultimodalZeroEval

Pricing & Availability

Input (per 1M tokens)

$2.50

Output (per 1M tokens)

$7.50

Max Input Tokens

1.0M

Max Output Tokens

65.5K

Supported Features

Function CallingStructured OutputCode ExecutionWeb SearchBatch InferenceFine-tuning

Benchmark Results

Model performance metrics across various tests and benchmarks

Programming

Programming skills tests

SWE-Bench Verified

• Self-reported

80.4%

Reasoning

Logical reasoning and analysis

GPQA

Diamond • Self-reported

92.4%

Other Tests

Specialized benchmarks

HMMT Feb 26

Self-reported by the model provider • Self-reported

97.0%

Kernel Bench L3

Fraction of problems faster than torch.compile • Self-reported

96.0%

BFCL-V4

• Self-reported

75.0%

Claw-Eval

Average • Self-reported

65.2%

CoWorkBench

• Self-reported

67.2%

CritPT

• Self-reported

11.4%

Finance Agent v2

• Verified

48.4%

Global PIQA

• Self-reported

91.4%

Humanity's Last Exam

• Self-reported

41.4%

IFBench

• Self-reported

79.1%

IFEval

Strict prompt • Self-reported

94.3%

IMO-AnswerBench

• Self-reported

90.0%

Include

• Self-reported

86.2%

LiveBench

2026-01-08 • Verified

74.3%

LiveCodeBench v6

• Self-reported

91.6%

MathArena Apex

• Self-reported

44.5%

MAXIFE

Accuracy on English and multilingual prompts across 23 settings • Self-reported

89.2%

MCP Atlas

Public set, Gemini 2.5 Pro judger • Self-reported

76.4%

MCP-Mark

GitHub MCP v0.30.3 • Self-reported

60.8%

MMLU-Pro

• Self-reported

89.6%

MMLU-ProX

Average accuracy across 29 languages • Self-reported

87.0%

MMLU-Redux

• Self-reported

95.0%

MMMLU

• Self-reported

90.3%

MRCR 128K (8-needle)

MRCR-v2 128K context subset with 8 needles • Self-reported

90.4%

NL2Repo

Claude Code scaffold • Self-reported

47.2%

NOVA-63

• Self-reported

59.0%

PolyMATH

• Self-reported

86.5%

QwenSVG

Internal SVG generation benchmark, BT/Elo rating • Self-reported

80.4%

QwenWebBench

QwenWebDev internal front-end code generation benchmark, BT/Elo rating • Self-reported

78.4%

QwenWorldBench

Open-ended 5-dimensional rubric judge grounded in real-environment feedback • Self-reported

57.3%

SciCode

• Self-reported

53.5%

SkillsBench

OpenCode, 78 tasks, avg of 5 runs • Self-reported

59.2%

SpreadSheetBench-v1

• Self-reported

87.0%

SuperGPQA

• Self-reported

73.6%

SWE-bench Multilingual

• Self-reported

78.3%

SWE-Bench Pro

• Self-reported

60.6%

Terminal-Bench 2.0

Terminus-2 harness • Self-reported

69.7%

VITA-Bench

Average subdomain score • Self-reported

47.9%

WMT24++

Average across 55 languages via XCOMET-XXL • Self-reported

85.8%

ZClawBench

QwenClawBench real-user-distribution Claw agent benchmark • Self-reported

64.3%

License & Metadata

License

proprietary

Announcement Date

May 19, 2026

Last Updated

August 27, 2026

Similar Models

All Models

Qwen3 Max

Alibaba

Best score:0.6 (GPQA)

Released:Dec 2025

Qwen3.5 122B A10B

Alibaba

122.0B

Best score:0.9 (GPQA)

Released:Mar 2026

Qwen3.5 35B A3B

Alibaba

35.0B

Best score:0.8 (GPQA)

Released:Mar 2026

Qwen3 Max Thinking

Alibaba

1.0T

Best score:0.9 (GPQA)

Released:Feb 2026

Price:$1.20/1M tokens

Qwen3.5 27B

Alibaba

27.0B

Best score:0.9 (GPQA)

Released:Mar 2026

Qwen3.6 Plus

Alibaba

Best score:0.9 (GPQA)

Released:Mar 2026

Price:$0.50/1M tokens

Qwen3.7-Plus

Alibaba

Best score:0.9 (GPQA)

Released:May 2026

Price:$0.32/1M tokens

MiMo-V2-Pro

Xiaomi

Best score:1.0 (TAU)

Released:Mar 2026

Recommendations are based on similarity of characteristics: developer organization, multimodality, parameter size, and benchmark performance. Choose a model to compare or go to the full catalog to browse all available AI models.