Qwen3.7-Max

Qwen's flagship model for the agent era with 1M context and long-horizon autonomy

Try Now

read docs

About model

Qwen3.7-Max is Qwen's flagship proprietary model built for the agent era, combining frontier reasoning with deep, generalizable agentic capabilities across coding, office automation, and long-horizon task execution. It leads on Terminal-Bench 2.0-Terminus (69.7) and achieves 92.4% GPQA Diamond, 80.4% SWE-Bench Verified, and 97.1% HMMT 2026 Feb. According to Qwen, the model maintained coherent execution across a ~35-hour autonomous session, generalizing across agent scaffolds without framework-specific tuning. Available on Together AI with a 1M token context window.

Terminal-Bench 2.0-Terminus

69.7

Leads the field on agentic terminal coding and execution

GPQA Diamond

92.40%

Top scientific and mathematical reasoning

Autonomous Execution

35hrs

10.0x kernel speedup on unseen hardware

Model key capabilities

Agentic Coding: 80.4% SWE-Bench Verified and 78.3% SWE-Multilingual.
Long-Horizon Autonomy: Maintained coherent execution across a Qwen-reported ~35-hour session, demonstrating a 10.0x kernel speedup on unseen hardware.
General Agent Workflows: Strong MCP tool orchestration (60.8% MCP-Mark, 76.4% MCP-Atlas) and office automation including spreadsheet work at 87.0% SpreadSheetBench-v1.
Reasoning and Instruction Following: 92.4% GPQA Diamond and 97.1% HMMT 2026 Feb for scientific and mathematical reasoning, 79.1% IFBench for instruction following.

Performance benchmarks

Model	GPQA Diamond	HLE	LiveCodeBench	MATH500	SWE-bench verified
Qwen3.7-Max			91.6		80.4
Related open-source models
Competitor closed-source models
Claude Opus 4.6	90.5%	34.2%			78.7%
OpenAI o3	83.3%	24.9%		99.2%	62.3%
OpenAI o1	76.8%			96.4%	48.9%
GPT-4o	49.2%	2.7%	32.3%	89.3%	31.0%

API usage

cURL
Python
Typescript

Endpoint:

Qwen/Qwen3.7-Max

curl -X POST "https://api.together.xyz/v1/chat/completions" \
  -H "Authorization: Bearer $TOGETHER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "Qwen/Qwen3.7-Max",
    "messages": [
      {
        "role": "user",
        "content": "What are some fun things to do in New York?"
      }
    ]
}'

from together import Together

client = Together()

response = client.chat.completions.create(
  model="Qwen/Qwen3.7-Max",
  messages=[
    {
      "role": "user",
      "content": "What are some fun things to do in New York?"
    }
  ]
)
print(response.choices[0].message.content)

import Together from 'together-ai';
const together = new Together();

const completion = await together.chat.completions.create({
  model: 'Qwen/Qwen3.7-Max',
  messages: [
    {
      role: 'user',
      content: 'What are some fun things to do in New York?'
     }
  ],
});

console.log(completion.choices[0].message.content);

Model card
Architecture Overview:
• Proprietary model with 1.0M token context window
• Text input and output; text-only

Training Methodology:
• Decoupled tasks, execution frameworks, and validators during training and employed cross-framework reinforcement learning to avoid shortcut overfitting to specific benchmarks

Performance Characteristics:
• Coding agents: 69.7 Terminal-Bench 2.0-Terminus, 80.4% SWE-Verified, 60.6% SWE-Pro, 78.3% SWE-Multilingual, 53.5% SciCode
• General agents: 60.8% MCP-Mark, 76.4% MCP-Atlas, 87.0% SpreadSheetBench-v1, 75.0% BFCL-V4
• Reasoning: 92.4% GPQA Diamond, 97.1% HMMT 2026 Feb, 41.4% HLE, 91.6% LiveCodeBench
• Instruction following: 79.1% IFBench, 94.3% IFEval
• Multilingual: 85.8% WMT24++, 90.3% MMMLU, 89.2% MAXIFE
• Long context: 90.4% MRCR-v2 128K
‍
Prompting
Together AI API Access:
• Access Qwen3.7-Max via Together AI APIs using the endpoint Qwen/Qwen3.7-Max
• Authenticate using your Together AI API key in request headers
• Available on Together AI with 1M token context
‍
Applications & use cases
Agentic Coding & Software Engineering:
• Repository-level reasoning with 80.4% SWE-Verified and 60.6% SWE-Pro
• Cross-language engineering with 78.3% SWE-Multilingual

Long-Horizon Autonomous Tasks:
• Sustained execution across multi-hour sessions, maintaining coherent strategy according to Qwen
• Kernel optimization, hardware profiling, and iterative technical tasks on unseen platforms

Office Automation & Productivity:
• Document generation, data analysis, and formatting through MCP and multi-agent orchestration
• 87.0% SpreadSheetBench-v1 for complex spreadsheet reasoning and automation

Reasoning & Multilingual Workflows:
• 92.4% GPQA Diamond and 97.1% HMMT 2026 Feb for scientific and mathematical problem solving
• 79.1% IFBench for precise instruction following in complex multi-step tasks
• 85.8% WMT24++ and 90.3% MMMLU for multilingual understanding and translation
‍

Related models

Model specifications

Model data

Model provider
Qwen
Type
Reasoning
Chat
Code
LLM
Main use cases
Reasoning
Features
Function Calling
JSON Mode
Deployment
Serverless
Endpoint
Qwen/Qwen3.7-Max
Context length
1M
Input price
$2.50 / 1M tokens
Output price
$7.50 / 1M tokens
Input modalities
Text
Output modalities
Text

Released
May 18, 2026
Category
Chat

Run in Playground

Quickstart docs

Deploy model

Qwen3.7-Max

About model

API usage

Model card

Prompting

Applications & use cases