- Alibaba has discontinued the free tier for its Qwen Code AI coding assistant, directing users to paid third-party providers for cloud access.
- This follows a similar move by Chinese AI lab MiniMax, which rewrote its licensing terms to require authorization for commercial use shortly after releasing its powerful M2.7 model.
- These shifts signal a strategic pivot away from freely accessible services as Chinese AI models, which achieved massive global adoption through open access, now face pressure to generate returns.
- The core Qwen models remain downloadable, but their most capable versions require significant local hardware to run.
Alibaba terminated its free cloud service for the Qwen Code AI programming tool today, a sudden pivot that removes a key free alternative to paid rivals like Claude Code. This decision emerged directly from the project’s GitHub repository, which now states the “Qwen OAuth free tier has been discontinued.”
The company also slashed the daily free request allowance from 1,000 to a mere 100. Consequently, users are now guided toward paid services such as Alibaba Cloud’s $50 monthly Coding Plan Pro, OpenRouter, or Fireworks AI for cloud-based access.
This move arrived just 48 hours after a parallel action by MiniMax. The Chinese lab released its high-performing 230 billion-parameter M2.7 model, then promptly revised its license to demand written permission for commercial applications, a significant departure from its initial “MIT-style” terms which imposed no such restrictions.
MiniMax defended the change by citing concerns about third-party Hosting providers distributing inferior versions under its brand. The company’s Head of Developer Relations argued that these providers degraded the user experience, leading people to conclude “MiniMax is mid.”
These strategic adjustments are not coincidental. According to the Financial Times, Alibaba’s own Qwen team has been steering toward more proprietary development recently. Meanwhile, Chinese open-source models saw explosive growth, rising from 1.2% of global usage in late 2024 to approximately 30% by the end of 2025, with Qwen surpassing Meta’s Llama as the world’s most deployed self-hosted model.
That widespread adoption was fundamentally built on freely available services. However, with tightening U.S. chip export controls and intense geopolitical competition, the economic rationale for offering powerful tools for free is becoming increasingly difficult to justify to investors.
✅ Follow BITNEWSBOT on Telegram, Facebook, LinkedIn, X.com, and Google News for instant updates.
