Qwen3.5 Plus 2026-02-15

Qwen3.5-Plus is Alibaba's hosted multimodal API model, offering the full capabilities of Qwen3.5-397B-A17B with production features including 1M context, built-in tools, and adaptive tool use.

Qwen3.5-Plus is the hosted version corresponding to Qwen3.5-397B-A17B with more production features, e.g., 1M context length by default, official built-in tools, and adaptive tool use. — Qwen Team

Overview

Released February 15, 2026 via Alibaba Cloud Model Studio, Qwen3.5-Plus provides managed access to Qwen3.5's unified vision-language capabilities. It extends the open-weight model with production-ready features: million-token context without configuration, official tool integrations, and adaptive tool invocation.

Key Features

1M context length by default (no YaRN configuration needed)
Built-in official tools: Search, code interpreter, and more
Adaptive tool use: Model decides when to invoke tools
Unified multimodal: Text, image, and video understanding
201 languages supported
Production infrastructure: Managed scaling and reliability

Production Advantages Over Open Weights

Feature	Qwen3.5-Plus (API)	Qwen3.5-397B-A17B (Self-hosted)
Context Length	1M default	262K native, 1M with config
Built-in Tools	Yes	Manual integration
Adaptive Tool Use	Yes	Manual implementation
Deployment	Managed	8+ GPUs required
Scaling	Automatic	Self-managed

When to Use Qwen3.5-Plus

Choose Qwen3.5-Plus when you need:

Million-token context for massive documents or codebases
Built-in tools without custom integration
Adaptive tool invocation (model decides when to search/execute)
Production-ready deployment without infrastructure
Multimodal capabilities (text, image, video)
Fast iteration without GPU provisioning

Choose Qwen3.5-397B-A17B (open weights) when you need:

Full control over deployment and fine-tuning
Data privacy with on-premise hosting
Cost optimization at high volume
Custom tool integrations

Choose Qwen3-Max-Thinking when you need:

Text-only workloads with maximum reasoning
Established Qwen3 ecosystem compatibility

Capabilities

Vision Understanding:

Document OCR in 32 languages
Chart and diagram analysis
Visual math problem solving
GUI operation for visual agents

Long Context:

Process entire codebases
Analyze lengthy documents
Multi-document synthesis
Long video understanding

Agent Tasks:

Autonomous tool selection
Multi-step planning and execution
Web browsing and search
Code execution

Role in Series

Qwen API models:

Qwen-Flash: Lowest latency
Qwen-Turbo: Fast, cost-effective
Qwen-Plus: Balanced performance
Qwen3-Max: Text-focused flagship
Qwen3.5-Plus: Unified multimodal flagship (this model)

Qwen: Qwen3.5 Plus 2026-02-15

Model Type