Model

Qwen3.7-Plus

CLOSED

Alibaba · Qwen3.7 family · released Jun 2, 2026

Cheaper, multimodal (vision + video input) sibling of Qwen3.7-Max released ~2 weeks later on the Bailian/Model Studio platform; adds GUI grounding and screen/mobile agent capability (ScreenSpot Pro, AndroidWorld) on top of the Max agent backbone. Text output only (no image/video generation). Preceded by a preview from ~May 14, 2026.

ReasoningCodingVisionFunction callingTool useAgentic

2332.0

Elo · rank #59

Parameters

Undisclosed

Active params

Undisclosed

Context

1M tokens

Architecture

MoE, exact expert configuration undisclosed; multimodal agent variant built on the Qwen3.7-Max backbone

License

Proprietary

Languages

—

API price (in/out)

$0.4 / $1.6

Modalities

text · vision · video

Benchmark results

Bar shows position within the tracked field; marker = field best

AIMEMath93.3%#29

best: GPT-5.2 · 100.0%

AndroidWorldAgents81.0%#4

best: Qwen3.8-Omni-Flash · 87.1%

Arena EloHuman preference1456#34

best: Claude Fable 5 · 1504

CharXivVision85.9%#5

best: Qwen3.8-Flash-Next · 90.6%

DeepSWECoding14.2%#26

best: Gemini 4 Argon · 77.9%

GPQA DiamondReasoning90.3%#32

best: GPT-6 Astra · 96.0%

Humanity's Last ExamReasoning34.7%#46

best: Claude Opus 5.5 · 67.7%

IFBenchReasoning79.1%#13

best: MiniMax M3 · 83.0%

IFEvalReasoning94.6%#4

best: Gemma 4 26B A4B · 98.5%

JobBenchAgents27.6%#11

best: Claude Opus 5 · 65.7%

LiveCodeBenchCoding89.6%#11

best: DeepSeek-V4-Pro (Think Max) · 93.5%

MathVisionVision90.3%#4

best: Seed 2.1 Pro · 92.6%

MMLU-ProKnowledge88.5%#7

best: Claude Fable 5 · 91.5%

MMLU-ReduxKnowledge94.5%#4

best: Qwen3.7-Max · 95.0%

MMMLUKnowledge89.0%#10

best: Gemini 3.1 Pro · 92.6%

MMMU-ProVision79.0%#15

best: Claude Opus 4.7 · 85.5%

NL2RepoCoding41.1%#8

best: DeepSeek-V4.1-Flash · 64.0%

OSWorld-VerifiedAgents73.3%#15

best: Claude Fable 5 · 85.0%

SuperGPQAReasoning71.4%#3

best: Qwen3.7-Max · 73.6%

SWE-bench ProCoding57.6%#28

best: Claude Opus 5.5 · 89.9%

SWE-bench VerifiedCoding77.7%#39

best: Claude Opus 5 · 97.0%

Terminal-Bench 2.0Coding70.3%#30

best: DeepSeek-V4.1-Flash · 90.6%

Video-MMEVision88.0%#2

best: Seed 2.1 Pro · 89.2%

WebArena-VerifiedAgents55.3%#2

best: Qwen3.8-27B · 64.8%

Run it locally

Closed weights — available via API only. No local deployment.

Input / M tok

$0.4

Output / M tok

$1.6

Qwen3.7 family

Elo progression across releases

Qwen3.7-MaxMay 20262361.2 Qwen3.7-PlusJun 20262332.0 Qwen3.7 FlashJul 20261000.0

Resources

DOCSwww.alibabacloud.com/help/en/model-studio/models APIAvailable via Alibaba API

API price $0.4/$1.6 · each benchmark row carries its own source badge (see methodology)