# Qwen 3.7 Max

Qwen 3.7 Max succeeds Qwen3.6-Max-Preview as the next flagship in Alibaba Cloud's Max tier, positioned as an agent foundation that sustains coherent reasoning across long-horizon tool-calling sessions.

- **Model ID:** `alibaba/qwen3.7-max`
- **Type:** chat
- **Providers:** alibaba
- **Context window:** 991,000
- **Maximum output tokens:** 64,000
- **Pricing:** $2.5/1M input tokens, $7.5/1M output tokens
- **Canonical page:** https://vercel.com/ai-gateway/models/qwen3.7-max

## Supported parameters

Detailed capability metadata has not been reported for this model.

## Example

```ts
import { streamText } from 'ai'

const result = streamText({
  model: 'alibaba/qwen3.7-max',
  prompt: 'Why is the sky blue?'
})
```

## About

Qwen 3.7 Max is the Max-tier release in the Qwen 3.7 generation, succeeding Qwen3.6-Max-Preview in Alibaba Cloud's closed-weight API line. The model is served through Alibaba Cloud with a context window of 991K tokens, which suits full-repository ingestion, long agent traces, and multi-document analysis without segmentation.

Alibaba Cloud describes Qwen 3.7 Max as designed as an agent foundation. The model targets coding agents that plan and act across many turns, office and productivity tasks that route work through multi-agent orchestration, and long-horizon autonomous execution where the model must maintain coherent reasoning across hundreds of sequential tool calls. Reported improvements over Qwen3.6-Max-Preview concentrate in frontend prototyping and complex multi-file engineering work.

Like other Max-tier entries, Qwen 3.7 Max supports tool calling and structured outputs, with extended-thinking mode available for high-difficulty reasoning, scientific computation, and expert-level queries. The thinking budget can be tuned per request to balance depth of reasoning against latency and token spend. Qwen 3.7 Max is text-only; for vision input, the sibling Qwen3.7-Plus is the multimodal entry in the 3.7 lineup.

You can integrate Qwen 3.7 Max through AI SDK, Chat Completions API, Responses API, Messages API, or other API formats, from TypeScript or Python.

## What to consider

Agent workflows that chain hundreds of tool calls produce high output-token volume. Use the AI Gateway cost dashboard to monitor per-session spend and tune the thinking budget before running production traffic at scale.

## When to use

### Best For

- Long-horizon coding agents that chain many tool calls and need to back out of dead ends without losing earlier context
- Multi-file software engineering work, including refactoring, diff editing, and frontend prototyping across a repository
- Office and productivity automation that routes work through multi-agent orchestration
- Expert-level reasoning tasks in scientific computation, mathematics, and structured analysis where extended thinking improves accuracy
- Long-context workloads that benefit from the window of 991K tokens, including full-codebase ingestion and document plus tool-trace inputs

### Consider Alternatives When

- Vision or multimodal input is required, in which case Qwen3.7-Plus is the multimodal entry in the 3.7 line
- Latency dominates and extended-thinking traces add unnecessary overhead, where a Plus or Flash-tier model serves users better
- Token budget per session is tightly constrained and the workload doesn't justify a flagship Max-tier model
- Visible chain-of-thought with built-in autonomous search or code execution is required, where Qwen3-Max-Thinking is a stronger fit

## Best for

- **Long-Horizon Coding Agents:** Sustained tool-calling sessions across many turns with planning, retries, and dead-end recovery
- **Multi-File Software Engineering:** Refactoring, diff editing, and frontend prototyping across a repository
- **Office Workflow Automation:** Routing productivity tasks through multi-agent orchestration
- **Expert Reasoning Tasks:** Scientific computation, mathematics, and structured analysis with extended-thinking mode
- **Repository Ingestion:** Long-context workloads using the window of 991K tokens for full codebases and tool traces

## Consider alternatives

- **Vision Or Multimodal Input:** Qwen3.7-Plus is the multimodal entry in the 3.7 line when image inputs are needed
- **Latency-Sensitive Pipelines:** A Plus or Flash-tier model serves users better when extended-thinking traces add unnecessary overhead
- **Strict Token Budgets:** A smaller model is a closer fit when per-session spend on a flagship Max model isn't justified
- **Built-In Autonomous Search:** Qwen3-Max-Thinking is a stronger fit when integrated search and code interpreter tools are required

## Frequently asked questions

### What is the relationship between Qwen 3.7 Max and Qwen3.6-Max-Preview?

Qwen 3.7 Max is the Max-tier flagship in the Qwen 3.7 generation, succeeding Qwen3.6-Max-Preview. Alibaba Cloud positions Qwen 3.7 Max as an agent foundation with improvements in long-horizon tool use, multi-file coding, and office workflow automation.

### What is the context window for Qwen 3.7 Max?

The context window is 991K tokens. This supports full-repository ingestion, long agent traces with hundreds of sequential tool calls, and multi-document analysis without segmentation.

### Can Qwen 3.7 Max accept image or file inputs?

Qwen 3.7 Max is text-only. For vision input within the 3.7 line, use Qwen3.7-Plus, which is the multimodal entry in the generation.

### Does Qwen 3.7 Max support tool calling?

Yes. Qwen 3.7 Max supports structured tool calling and is tuned for agent workflows that chain many sequential tool invocations across long-horizon sessions.

### How does extended-thinking mode work on Qwen 3.7 Max?

Extended-thinking mode generates an internal reasoning trace before producing the final answer, which improves accuracy on high-difficulty logical reasoning, scientific computation, and expert-level queries. The thinking budget is tunable per request to balance depth against latency and token spend.

### How do I access Qwen 3.7 Max through AI Gateway?

Authenticate with an AI Gateway API key or OIDC token and reference ``alibaba/qwen3.7-max`` as the model. You can call Qwen 3.7 Max through AI SDK, Chat Completions API, Responses API, Messages API, or other API formats, from TypeScript or Python.

### Does Qwen 3.7 Max support zero data retention?

Yes, Zero Data Retention is available for this model. Zero Data Retention is offered on a per-provider basis. See https://vercel.com/docs/ai-gateway/capabilities/zdr for details.

### Where can I see live latency and cost data for Qwen 3.7 Max?

This page shows live throughput, time-to-first-token, and pricing metrics for Qwen 3.7 Max measured across real AI Gateway traffic.

## Links

- [Model page](https://vercel.com/ai-gateway/models/qwen3.7-max)
- [AI Gateway documentation](https://vercel.com/docs/ai-gateway)
- [Provider model documentation](https://modelstudio.console.alibabacloud.com/ap-southeast-1?tab=doc#/doc/?type=model&url=2840914_2&modelId=qwen3.7-max&serviceSite=international)


If you encounter incorrect information, can’t find what you need, need a capability Vercel doesn’t currently offer, or have any other feedback to report, please do it by following the instructions at https://vercel.com/feedback.md.