AI Tools Navigator
Back to AI information
Qwen3-Max-Instruct/Thinking is now available: Coding and Agent capabilities are significantly enhanced

Qwen3-Max-Instruct/Thinking is now available: Coding and Agent capabilities are significantly enhanced

AI information • Admin • • 98 views

Qwen has announced the launch of its next-generation flagship model, Qwen3-Max , officially described as "no preview, ready for immediate use." The model is available in two formats: Instruct and Thinking . The former focuses on general-purpose implementation and engineering applications, while the latter enhances tool usage and complex reasoning. According to the release materials, Qwen3-Max-Instruct compares favorably with leading models on multiple benchmarks, including SWE-Bench, Tau2-Bench, SuperGPQA, LiveCodeBench, and AIME25 . Qwen3-Max-Thinking, when used in "heavy" mode and combined with tool invocation, claims to achieve near-perfect performance on key benchmarks.

Alibaba Cloud Model Studio has launched CODE INLINE 0 and snapshot versions, offering large context, tiered billing, and support for context caching. The official blog states that this series builds on the continuous expansion of larger data, pre-training, and reinforcement learning (RL), targeting coding, agent workflows, and long document scenarios. Actual performance, pricing, and quotas are subject to the model page and console.

Frequently Asked Questions

Q: Where can I experience and call it?

A: Qwen Chat provides an online experience; Alibaba Cloud Model Studio opens the qwen3-max API and snapshots.

Q: What is the difference between Instruct and Thinking?

A: Instruct is aimed at general use and engineering implementation; Thinking combines tool usage and "heavy" mode, focusing on deep reasoning.

Q: What are the public indicators?

A: The official results are listed as leading/benchmarking on benchmarks such as SWE-Bench, Tau2-Bench, SuperGPQA, LiveCodeBench, and AIME25 . Please refer to the blog and model page for details.

Q: How does context relate to billing?

A: The Model Studio page displays 256K level contexts with tiered pricing, and supports context caching and snapshot versions.

Q: Is the weight open source?

A: This is a product release and is available online. Whether it will be open source will be subject to subsequent announcements.

Qwen3-Max released Qwen3-Max flagship model Qwen3-MaxInstruct Qwen3-MaxThinking Qwen3-Maxheavy Mode Qwen3-Max tool call Qwen3-Max Complex Reasoning Qwen3-Max project launched Qwen3-Max encoding capabilities Qwen3-MaxSWE-Bench Qwen3-MaxTau2-Bench Qwen3-MaxSuperGPQA Qwen3-MaxLiveCodeBench Qwen3-MaxAIME25 Qwen3-Max benchmarks top Qwen3-Max can be used directly without preview Qwen3-Max long context 256K Qwen3-Max Context Cache Qwen3-Max snapshot version Qwen3-Max Tiered Billing Qwen3-MaxModelStudio Qwen3-MaxQwenChat Experience Qwen3-Max API Call Qwen3-Max Agent Workflow Qwen3-Max long document scenario Qwen3-Max Thought Chain Qwen3-Max Reinforcement Learning RL Qwen3-Max large-scale pre-training Qwen3-Max tool collaboration Qwen3-Max Code Review Qwen3-Max multi-step repair Qwen3-Max Terminal Mission Qwen3-Max context window Qwen3-Max Price Quota Qwen3-Max model page information Qwen3-Max console activated Qwen3-Max benchmark scores Qwen3-Max Enterprise Launch Qwen3-Max Developer Guide Qwen3-Max Best Practices Qwen3-Max and Qwen3-VL work together Qwen3-Max vs. GPT-5 Qwen3-MaxHeavy nearly full score Qwen3-Max tool chain integration Qwen3-Max Code Proxy Qwen3-Max long session stability Qwen3-Max context cache savings Qwen3-Max multimodal coordination Qwen3-Max model upgrade Qwen3-Max FAQ

Related Articles

Qwen3-VL Released: Flagship 235B Model Open Source, Instruction/Thinking Versions Available

Qwen3-VL Released: Flagship 235B Model Open Source, Instruction/Thinking Versions Available

Tongyi Qianwen has launched the next-generation visual language model, the Qwen3-VL . The flagship Q...

Chrome launches DevTools MCP server: AI can run performance tracking, inspect DOM and real-time debugging

Chrome launches DevTools MCP server: AI can run performance tracking, inspect DOM and real-time debugging

The Chrome team announced the public preview of the DevTools MCP (Model Context Protocol) server . T...

MWC Shanghai hosts the robot penalty shootout: embodied intelligence moves into the public exam venue

MWC Shanghai hosts the robot penalty shootout: embodied intelligence moves into the public exam venue

On June 1, 2026, IT Home reported that GSMA officially announced today that the "Humanoid Robot Pena...

Codex supports Windows control: AI programming agents begin cross-platform collaboration

Codex supports Windows control: AI programming agents begin cross-platform collaboration

On June 1, 2026, the ChatGPT Release Notes from the OpenAI Help Center still show as recent updates....

Recommended Tools

More