Qwen3.6-Plus

by Alibaba

Qwen 3.6 Plus is Alibaba's flagship agent-optimized language model, built on a sparse mixture-of-experts architecture. With a 1 million token context window, it can hold large codebases, technical documentation, or extended conversation histories in a single request, reducing the need for external retrieval in many long-document scenarios. The model accepts vision input alongside text, natively handles function calling and tool orchestration, and offers configurable chain-of-thought reasoning. It is built for agentic coding, complex document and visual understanding, and tasks that require sustained reasoning over large contexts, including frontend and visual coding from UI screenshots. Qwen 3.6 Plus is designed for real-world agent workflows and supports the Model Context Protocol (MCP) along with common coding assistants. A preserve_thinking option carries reasoning content across conversation turns, which helps maintain consistency in multi-step agentic tasks.

Key info

Input
Output
Features
Context window
1M
Max output
66K

Benchmarks

Independent benchmark scores: the Intelligence Index and the individual evaluations behind it, grouped by skill.

Global rank#103 of 508 LLMs
TierEfficient
Output speed56 tok/s
Intelligence Index (previous version)27.0
Agentic & knowledge work
GDPval
25%
Terminal-Bench Hard
44%
τ²-Bench Telecom
98%
Science & reasoning
Humanity's Last Exam
28%
GPQA Diamond
88%
Long-context reasoning
78%
Knowledge
MMMU-Pro
78%
Math & instruction following
IFBench
75%

Available routes

No routes currently available — Qwen3.6-Plus isn't routed through the Opper gateway right now. It may return.

Contact us about this model →

Available models from Alibaba

Start building with 700+ models

One API key for every major provider, up and running in minutes.

Get startedView Documentation