AI Briefing
KO

QwQ-Max-Preview Released

·2025.02.25 03:00

Key point

Qwen has released **QwQ-Max-Preview**, which enhances reasoning and coding performance, and has previewed future app launches and open-sourcing of small models.

Details

The latest model in the Qwen series, QwQ-Max-Preview, has been released. Built on Qwen2.5-Max, this preview model is designed to deliver excellent performance in deep reasoning, math, coding, general domains, and Agent-related workflows.

This version is a stage for previewing the model's enhanced capabilities before its official release. There are plans to officially open-source QwQ-Max and Qwen2.5-Max under the Apache 2.0 license in the future.

The future roadmap includes the following key plans:

  • Qwen Chat APP launch: Providing an intuitive interface that allows general users to perform real-time reasoning, code generation, and logical problem-solving without technical expertise.
  • Open-sourcing small models: Releasing lightweight reasoning models such as QwQ-32B so that developers and researchers can use them locally while maintaining low latency and privacy protection.
llmopen-sourceqwenai-agentsqwq-max-previewdeep-reasoning
Read originalShare on Threads

This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.

Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.

Related

RedditMica v0.1 4B Released: Open-Source Decision Model for Agent Loops Trained for Under $30RedditUkisAI Releases Swift Series: Efficient Qwen-Based LLMs with Reduced Thinking TokensQwenQwen2.5-1M: Qwen Releases Models Supporting Up to 1 Million Token Context LengthPyTorchKRvLLM Releases 'Decision 1.0', AI Models Specialized for Decision-MakingGeekNewsWhiteboard Launches Open-Source VS Code Fork for Human-AI Collaborative Software DesignHacker NewsOpenComputer Hosts 'Ship Video' Agent Using Claude Opus 5.5 to Generate Explainer Videos from Code