Qwen 3.7 Max
QwenFlagship
ThinkingTool UseStructured Output
About this model
Qwen3.7-Max is the flagship model in Alibaba's Qwen3.7 series. It supports text input and output and is designed for agent-centric workloads, with particular strengths in coding, office and productivity tasks,...
Performance Tier
Flagship
Qwen 3.7 Max is a flagship model from Qwen : the most capable in their lineup.
Best-in-class model from this provider. Highest performance across benchmarks, ideal for demanding tasks.
Pricing
This model is included in Elosia plansModerate
Moderate cost. A balanced choice for regular use without constant cap watching.
| Type | per 1M tokens |
|---|---|
| Input (prompt) | $2.50 |
| Output (completion) | $7.50 |
| Cache write | $3.13 |
Capabilities
Context Length1.0M
Max Output Tokens66K
TokenizerQwen
Inputtext
Outputtext
Release DateMay 21, 2026
Benchmarks
General Intelligence
MMLU
Not reported
GPQA Diamond
92.4%
Mathematics
MATH-500
Not reported
Programming
HumanEval
Not reported
SWE-bench Verified
80.4%
SWE-bench Multilingual
78.3%
Reasoning
IFEval
Not reported
Humanity's Last Exam
41.4%
Agentic
SWE-bench Pro
60.6%
Terminal-Bench 2.0
69.7%
Recommended Use Cases
CodingAnalysisResearchGeneral ChatTranslation
Strengths
- Alibaba's most capable agent model — leads SWE-bench Verified (80.4%) on par with Claude Opus 4.6
- Top-tier scientific reasoning (GPQA Diamond 92.4%) — ahead of Claude Opus 4.6 (91.3%)
- Strong agentic coding across languages (SWE-bench Multilingual 78.3%, SWE-bench Pro 60.6%)
- Leads Terminal-Bench 2.0 (69.7%) for long-horizon command-line workflows
- Built-in thinking mode (enable_thinking / preserve_thinking) for chain-of-thought reasoning
Limitations
- Alibaba's evaluation focuses on agentic and coding benchmarks — MMLU, MATH-500, HumanEval, AIME and LiveCodeBench are not part of the reported suite
- Native context advertised at 256K tokens (1M only mentioned for some configurations)
- Premium pricing among Qwen models ($2.50 / $7.50 per M tokens)
Resources
This model may use your data for training
