Skip to content
Qwen3.7 Max logo

Qwen3.7 Max

Alibaba

Key facts · verified Oct 30, 2026

Qwen3.7 Max is an Alibaba model released on 20 May 2026. Its API price is $2.50 per million input tokens and $7.50 per million output tokens ($0.5 implicit cached input (Alibaba Cloud Model Studio, International)). Its context window is 1M tokens. It ranks #42 on LMArena's text leaderboard (Elo 1474, read 1 Nov 2026). Source: alibabacloud.com.

Qwen3.7 Max (qwen3.7-max) is a proprietary model from Alibaba on Alibaba Cloud Model Studio, which Alibaba describes as the largest and most capable model in the Qwen3.7 series, built for agent-centric use, with a 1M-token context window and up to 131,072 output tokens. Alibaba states the qwen3.7-max alias is functionally equivalent to the qwen3.7-max-2026-05-20 snapshot; a separate qwen3.7-max-2026-06-08 snapshot adds image and video input. Alibaba lists international pricing of $2.50 per million input tokens and $7.50 per million output tokens. 1474 Elo (rank 42)

Pricing & specs checkedSource: vendor documentation · checked Oct 30, 2026LMArena: 1474 Elo (rank 42) · source · retrieved Nov 1, 2026
Pricing

$2.50 / $7.50 per million tokens

Context

1M tokens

LMArena Elo

1474

LMArena rank

#42

Key Features

Agentic Coding Office & Productivity Multimodal Function Calling Context Caching

Full Specifications

$2.50 / $7.50 per million tokens
Identity
DeveloperAlibaba
ReleasedMay 2026
StatusGA
LicenceProprietary
Self-hostableNo
Cost
Blended $/1M tokensinput × 0.75 + output × 0.25$3.750 / 1M tokens
Input price$2.500 / 1M tokens
Output price$7.500 / 1M tokens
Cached input$0.5 / 1M tokens (implicit cache; $3.125 explicit-cache write, $0.25 explicit-cache read; International)
Batch discountNot offered in the International and Global regions; China (Beijing) batch file $0.825 / $2.475 per 1M tokens
Free tierNot offered
Capacity
Context window1M tokens
Max output131,072 tokens
Long-context surchargeNone (flat rate across the 1M window)
Capability
Vision inNo - Alibaba lists Text as the only input modality for qwen3.7-max in every region
Audio inNo
Function callingYes
Structured outputYes
Extended reasoningYes
Web searchYes
Code executionYes
Access
APIYes
Chat appYes
Cloud marketplacesNot offered
Fine-tuningNot offered
Measured quality
LMArena Elo1474 (checked Nov 2026)
LMArena rankRank 42
GPQA Diamond91
SWE-bench Verified77
Elo per dollarLMArena Elo ÷ blended $/1M tokens393