Industry Insights 10 min read

Is the Era of Commercial‑Ready Chinese Open‑Source LLMs Ending? MiniMax M2.7 License Update

The MiniMax M2.7 model switched its open‑source license to forbid commercial use, igniting a heated debate about what constitutes commercial activity, prompting a community clarification that self‑hosted coding remains free, and leading to a revised license that explicitly permits personal, academic, and non‑profit uses while highlighting broader market pressures from cloud providers that are reshaping the open‑source LLM ecosystem.

Old Zhang's AI Learning
Old Zhang's AI Learning
Old Zhang's AI Learning
Is the Era of Commercial‑Ready Chinese Open‑Source LLMs Ending? MiniMax M2.7 License Update

Free commercially usable large models

In the past two years DeepSeek, Qwen and GLM were released under MIT or Apache‑2.0 licenses, which allow unrestricted commercial use.

Recent events show a shift: open‑source models are increasingly restricting commercial use.

MiniMax M2.7 – license explicitly bans commercial use without written permission.

Kimi K2.5 – adds a commercial clause that requires prominent labeling when monthly active users exceed 100 million or monthly revenue exceeds $20 million.

Qwen 3.6‑Plus – closed‑source, providing only API access and no model weights.

MiniMax M2.7 license details

The license claims to be “MIT‑style” but adds three restrictive clauses:

Free use is limited to non‑commercial purposes such as personal research, learning or local experimentation.

Commercial use requires prior written authorization from MiniMax and a mandatory “Built with MiniMax M2.7” attribution.

“Commercial use” is defined broadly to include paid products or services, commercial API usage, and any deployment of fine‑tuned or otherwise modified versions.

Ollama offered M2.7 as a cloud service only after obtaining a commercial license.

Controversy: does self‑hosted coding count as commercial?

"Commercial Use" means any use of the Software or any derivative work thereof that is primarily intended for commercial advantage or monetary compensation, which includes, without limitation: (i) offering products or services to third parties for a fee… (ii) the commercial use of APIs… (iii) the deployment or provision of the Software or its derivatives that have been subjected to post‑training, fine‑tuning, instruction‑tuning, or any other form of modification, for any commercial purpose.

Interpreted literally, developers could not use a self‑hosted M2.7 instance to write code for a commercial application.

After community discussion, MiniMax developer‑relations lead Ryan Lee clarified that self‑hosted usage for coding is “absolutely allowed and free”.

License revision

MiniMax updated the LICENSE on HuggingFace, adding a fifth clause “Permitted Free Uses”:

5. Permitted Free Uses. The following uses are expressly permitted free of charge: (a) personal use, including self‑hosted deployment for coding, development of applications, agents, tools, integrations, research, experimentation, or other personal purposes; (b) use by non‑profit organizations, academic institutions, and researchers for non‑commercial research or educational purposes; (c) modification of the Software solely for the uses described in (a) or (b) above.

Thus self‑hosted coding, personal development, and non‑profit or academic research remain free.

Kimi K2.5 commercial clause

Kimi’s open‑source license adds a condition: when monthly active users exceed 100 million or monthly revenue exceeds $20 million, the product must prominently display “Kimi K2.5”. The threshold is high enough that most startups will not be affected.

Why the shift?

Cloud providers (Volcano Engine, Alibaba Cloud, Baidu Cloud) launched low‑price “Coding Plans” for open‑source models, dramatically undercutting the traditional “open‑source → API monetization” model.

Open‑source model free commercial use
→ Third‑party wrappers sell cheap API / cloud price war
→ Original model vendor’s API business is cannibalized
→ Brand reputation suffers from low‑quality deployments
→ Vendors restrict commercial use or go closed‑source

Model developers invest tens of millions to billions in training, yet see their models turned into low‑margin API products for cloud operators, while poor third‑party deployments damage their brand.

Emerging ecosystem tiers

Flagship closed‑source – closed‑source API, used in enterprise‑grade paid scenarios.

Limited open‑source – modified MIT or conditional commercial license, suitable for personal research or authorized commercial use.

Fully open‑source – Apache 2.0 or MIT license, allowing free commercial use for non‑flagship models.

AI industryMiniMaxopen-source modelsKimicommercial useLLM licensing
Old Zhang's AI Learning
Written by

Old Zhang's AI Learning

AI practitioner specializing in large-model evaluation and on-premise deployment, agents, AI programming, Vibe Coding, general AI, and broader tech trends, with daily original technical articles.

0 followers
Reader feedback

How this landed with the community

Sign in to like

Rate this article

Was this worth your time?

Sign in to rate
Discussion

0 Comments

Thoughtful readers leave field notes, pushback, and hard-won operational detail here.