Skip to main content
Alibaba

Qwen3.7 Max

Qwen3.7 Max is a large language model from Alibaba, positioned in the Max tier of the company's Qwen model family.

Input from
$1.25 / 1M tokens
across 7 providers

API Pricing

Cheapest on Novita AI — 42% below avg
ProviderInput / 1MOutput / 1MCached / 1M
$1.25$3.75$0.250
$1.48$4.42$0.295
$1.50$4.50-
$2.00$6.00$0.200
$2.50$7.50$0.500
$2.50$7.50$0.250
$3.75$11.25-

Prices updated daily. Last check: Sep 30, 2026

Qwen3.7 Max pricing by provider

Input, output, and batch rates, plus alternatives, for one provider at a time.

Performance & Benchmarks

Source: Artificial Analysis →
Intelligence
29.5 / 100
Coding
66.0 / 100

Reasoning & Knowledge

  • GPQA Diamond92.3%
  • Humanity's Last Exam40.5%

Coding

  • SciCode49.5%

Agentic & Tool Use

  • Terminal-Bench Hard50.8%
  • Terminal-Bench v2.174.5%
  • τ²-bench94.7%
  • τ-bench Banking11.8%

Instruction & Long Context

  • IFBench80.5%
  • Long-Context Reasoning79.0%

Benchmarks measured Sep 2026. Scores are independent evaluations, not vendor-reported.

Model Details

General

Creator
Alibaba
Modalities
Text

Capabilities

Tool Calling
No
Open Source
No

Strengths & Limitations

Strengths

  • Max-tier positioning within the Qwen family, the tier Alibaba reserves for its higher-capability general-purpose models
  • Part of the widely deployed Qwen series, which has a large ecosystem of tooling, integrations and community documentation
  • Qwen models are generally trained with strong Chinese-language coverage in addition to English, making them a common choice for bilingual workloads
  • Provides an alternative to US-based model providers for teams with sourcing or regional deployment requirements
  • Sits in a family with multiple tiers (Max, Plus, Flash), so workloads can be routed down to cheaper siblings without changing prompt format or provider

Limitations

  • We do not have a verified context window length on record for this model
  • No published benchmark scores for Qwen3.7 Max are recorded in our database, so capability comparisons must rely on vendor documentation
  • Throughput and time-to-first-token measurements from Artificial Analysis are not yet populated for this model
  • Input modality support (for example whether images are accepted) is not confirmed in our data
  • Provider availability for specific Qwen point releases can be narrower than for Alibaba's more widely mirrored versions

Key Features

•Max-tier model in Alibaba's Qwen3 generation lineage
•Available through inference API providers tracked in the pricing table on this page
•General-purpose text generation, reasoning and instruction following
•Bilingual English and Chinese usage typical of the Qwen series
•Sibling tiers (Plus, Flash, Turbo) available for cost-sensitive routing within the same family

About Qwen3.7 Max

Qwen3.7 Max is a large language model developed by Alibaba as part of its Qwen series. Within Qwen naming conventions, the "Max" designation marks the higher-capability tier of a generation, sitting above the smaller Plus, Flash and Turbo variants that Alibaba typically ships alongside it. The 3.7 version number places it in the Qwen3 generation lineage, as an iteration on the earlier Qwen3 and Qwen3 Max releases. Our database currently holds limited verified specifications for Qwen3.7 Max. We do not track a confirmed context window length, modality list, or published benchmark scores for this model, and the throughput and latency figures we have from Artificial Analysis are not yet populated. Rather than estimate, we list only what is confirmed: the creator (Alibaba) and the model's Max tier position in the Qwen family. Readers evaluating it for production should confirm context length, input modalities, and tool-calling support against Alibaba's own model documentation for the specific endpoint they plan to call. In practice, Max-tier Qwen models are the variants Alibaba markets for its most demanding workloads, and they are commonly compared against other large general-purpose models when teams want an alternative to US-based providers or want strong Chinese-language handling alongside English. Because provider availability for a given Qwen release varies — Alibaba's own Model Studio, plus third-party inference hosts — the practical decision often comes down to which providers serve this specific version and at what rates. The pricing table on this page lists the providers we currently track for Qwen3.7 Max.

Common Use Cases

As a Max-tier Qwen model, Qwen3.7 Max is aimed at workloads where a team wants Alibaba's stronger general-purpose option rather than a lightweight variant: multi-step reasoning, longer-form drafting and summarization, code generation and review, and assistant or agent backends where response quality matters more than per-token cost. It is a frequent candidate for products serving Chinese-language users, or bilingual Chinese/English content pipelines, given the Qwen series' training emphasis. For high-volume, low-complexity work such as classification, tagging or short extraction, the smaller Qwen tiers are usually the more economical fit. Because we do not have a confirmed context window or modality list for this release, verify those specifics against Alibaba's documentation before committing to long-document or image-input use cases.

Frequently Asked Questions

How much does Qwen3.7 Max cost to run?

Pricing varies by provider and by pricing type — input tokens, output tokens, and any cached-input or batch rates are all billed differently, and providers change rates frequently. Check the pricing table on this page for the current rates from each provider we track for Qwen3.7 Max.

What is Qwen3.7 Max best used for?

It suits general-purpose tasks where a higher-capability model is warranted: multi-step reasoning, code generation and review, long-form drafting and summarization, and assistant or agent backends. Its Qwen lineage also makes it a common pick for Chinese-language and bilingual Chinese/English applications. Route simpler high-volume tasks to smaller Qwen tiers instead.

How does Qwen3.7 Max differ from other Qwen tiers?

Alibaba uses Max for the higher-capability tier of a Qwen generation, with Plus, Flash and Turbo variants positioned below it for lower cost and faster responses. Qwen3.7 Max is therefore the option to reach for when task difficulty is the constraint, while the smaller tiers are better suited to throughput-driven workloads.

What is Qwen3.7 Max's context window?

We do not have a verified context window figure for Qwen3.7 Max in our database. Because Qwen releases differ on this point and individual providers sometimes cap context below the model's maximum, confirm the limit with Alibaba's model documentation and with the specific provider endpoint you plan to use.

Who created Qwen3.7 Max?

Qwen3.7 Max was developed by Alibaba as part of its Qwen family of large language models.