"Much better value for money! M2 and K2 are 'highly popular,' as China's open-source large models challenge 'AI programming.'"

```

Chinese AI companies are competing for a rapidly growing share of the AI programming market by offering prices far lower than OpenAI, Anthropic, and Google.

At the end of October, Shanghai-based AI startup MiniMax launched its open-source large model M2, primarily focused on programming and AI agents. M2's pricing is only 8% that of Anthropic's Claude Sonnet model.

Additionally, Beijing-based Moonshot AI's Kimi K2 model, released in September, has attracted attention. Venture capitalist Chamath Palihapitiya revealed last month that his firm’s employees have shifted much of their work to this open-source model, saying its "performance is stronger and it is much cheaper than OpenAI and Anthropic."

As Chinese models continue to prove their competitiveness in both performance and price, they not only open up vital overseas revenue sources for domestic AI firms but also secure a position in the fast-growing AI programming market.

Significant Price Advantage, Comparable Performance

The core appeal of Chinese AI models in the market lies in their outstanding “value for money.”

Take MiniMax’s M2 as an example: its cost per million output tokens is $1.20, while Anthropic’s Claude Sonnet model is as high as $15. This huge price difference has quickly made M2 a popular choice among developers.

Low cost does not come at the expense of performance. According to user voting on the AI model comparison platform LMArena, M2 ranks fourth in web development capabilities. On the AI model intelligence ranking compiled by Artificial Analysis, M2 ranks fifth.

Its popularity is also reflected in real-world usage: on OpenRouter, a platform offering hundreds of models to developers, M2 is currently the fourth most popular model by token usage, behind xAI, Anthropic, and Google.

Numerous startups developing coding assistant tools, including Cline, Kilo Code, and Roo Code, have already integrated M2 into their services as an option for users.

Favored by Tech Executives and Investors, Chinese Models Earn Endorsements

Besides the developer community’s warm reception, Chinese AI coding models have also gained recognition from American tech executives and venture capitalists, providing strong endorsements for their reliability.

Guillermo Rauch, CEO of US cloud platform Vercel, called Zhipu AI’s GLM-4.6 model, released in September, “an astonishingly excellent model” in a post on X.

Vercel has partnered with the company—founded by AI researchers from Tsinghua University—to offer the model to its users at the lowest price. GLM-4.6, which is mainly designed for coding, has shown strong performance on multiple rankings.

Meanwhile, venture capitalist Chamath Palihapitiya revealed on a podcast last month that his firm had migrated a substantial amount of workload to Moonshot AI’s open-source model Kimi K2.

He stated that Kimi K2 “performs much better and, frankly, is far cheaper than OpenAI and Anthropic.”

Moonshot AI released an updated version of Kimi K2 in September, further enhancing its coding capabilities.

Seeking Overseas Growth: Both Challenges and Opportunities

For Chinese startups such as DeepSeek, MiniMax, Moonshot AI, and Zhipu AI, attracting overseas paying users is vital.

Because corporate customers tend to be cautious about paying for subscriptions, AI companies face challenges generating revenue at home. Therefore, charging overseas clients through application programming interfaces (APIs) has become a key path to growth.

And with the overall expansion of the global AI coding market, demand for cheaper options with “good enough” performance will persist.

This presents a huge market opportunity for Chinese AI models, while also creating potential price war pressure for US AI companies relying on AI Agent businesses for revenue growth.

Risk Warning and DisclaimerThe market involves risks, and investment should be approached prudently. This article does not constitute personal investment advice and does not take into account individual users' special investment goals, financial situations, or needs. Users should consider whether any opinions, viewpoints, or conclusions herein are suitable to their particular circumstances. Investing based on this article is at your own risk. ```