Qwen3.7 Max

May 2026
Qwen3.7 Max
Compare in
on
$0.93per task?
$1.48 / $4.43per 1M in / outcache $0.29
202 tok/sfirst token in 2.2 s?
1Mcontext

Rankingin thinking mode?

Full ranking →
1Opus 5.51,866.32Muse Wonder Full1,857.13Sonnet 5.51,837.84Muse Wonder1,770.35Fable 5.11,763.6
···
57Inkling Small1,132.758Qwen3.7 Max1,128.259MiMo v2.5 Pro1,120.9

Strong at

S
Tool calling#21 of 242Science#19 of 312Instructions#4 of 245
A
Reasoning#45 of 336Long context#59 of 314Knowledge#36 of 320
B
Coding#54 of 134Agents & work tasks#58 of 177

Alternatives?

ModelScoreCostPer taskWhat changes
CheaperDeepSeek V4.1 Flash 1017maxOpen+55%-72%$0.26Pass back reasoning_content · Chat Completions+4BetterGPT-6 Lunamedium+13%-98%$0.017Function calling only with none · Chat Completions+3CheaperMiMo v2.6 ProthinkingOpen+50%-86%$0.13Pass back reasoning_content+3CheaperGLM 5.3 FlashmaxOpen+46%-70%$0.28Defaults to max · Self-hosted+4CheaperQwen3.8 FlashthinkingOpen+45%-60%$0.37Context: 1M → 262K+6

History

7 versions
1,2001,4001,600Jul 26Release dateGDPval-AA 2.13.7 Max3.8 Max3.8 Max (0902
1,2001,4001,600Jul 26Release dateGDPval-AA 2.13.7 Max3.8 Max

Benchmarks

in Agents & Tool Use
Terminal-Bench 2.174.5#31
Terminal-Bench 4.01.5#54
Terminal-Bench Hard (version unstated)50.8#12
GDPval-AA (work tasks) 2.11,128.2#58
AA Analyst Agent18.8#33
AA-LCR (long context) 1.179#59
τ²-bench · AA94.7#21
τ³-Banking · AA11.8#85
ITBench SRE42.5#14
AutomationBench-AA23#58
EnterpriseOps-Gym-AA45#11
AA-Briefcase 1.1 · Elo916.2#65
SWE-bench Pro (edition unstated)lab60.6
DeepSWE 1.1lab21.6
DeepSWE (edition unstated)lab18

FAQ

How much does Qwen3.7 Max cost?›

$1.48 per 1M input tokens and $4.43 per 1M output tokens; cached input $0.29 per 1M; cheapest via Novita at $1.25 / $3.75. A typical benchmark task costs about $0.93 at default effort.

Is Qwen3.7 Max free to use?›

There is no free API; the cheapest paid price is $1.25 per 1M input tokens.

How good is Qwen3.7 Max, and what is it best at?›

#45 of 336 in Chat & Knowledge, #54 of 134 in Coding, #58 of 135 in Agents & Tool Use, #39 of 44 in Writing & Language, #66 of 135 in Vision & Documents. Strongest at: Tool calling, Reasoning, Science, Long context, Knowledge, Instructions.

Is Qwen3.7 Max better than Inkling Small?›

On GDPval-AA (work tasks) 2.1, Qwen3.7 Max scores 1,128.2 vs 1,132.7 — lower. It is 942% more expensive per task.

How can I use Qwen3.7 Max?›

Through the API of Alibaba, DeepInfra, Novita.

What is the context window of Qwen3.7 Max?›

1M tokens, with up to 131K output tokens.

When was Qwen3.7 Max released? Is there a newer version?›

Released 2026-05-21. It has been superseded by Qwen3.8 Max.

Is Qwen3.7 Max open source?›

No, Qwen3.7 Max is a closed model available through an API. The closest open-weights option here is DeepSeek V4.1 Flash 1017.

Can Qwen3.7 Max read images?›

No, it accepts text only.

What is Qwen3.7 Max?›

Qwen3.7 Max is a closed language model from Alibaba (Qwen), released May 2026, the successor to Qwen3.6 Max Preview.

What are the best alternatives to Qwen3.7 Max?›

DeepSeek V4.1 Flash 1017 (cheaper), GPT-6 Luna (better).