Wealthari
Wednesday, Mar 4, 2026
  • Legal & Support :
  • About Us
  • Privacy Policy
  • Terms of Service
  • Contact Us
  • Home
  • Business
  • Technology
  • Leadership
  • Reviews
    • Business Services
      • LLC Comparisons
      • Accounting Software
      • Hosting Services
      • Insurance Solutions
    • Productivity Software
      • Asana vs Trello
      • Cloud Storage Apps
      • Time Tracking Tools
    • Learning Platforms
      • Online Certifications
      • Skillshare vs Udemy
    • Marketing ToolsNew
Reading: Chinese AI Models Capture Majority of OpenRouter Token Volume as MiniMax M2.5 Surges to the Top
Newsletter
Wealthari
  • Home
  • Business
  • Technology
  • Leadership
  • Reviews
    • Business Services
      • LLC Comparisons
      • Accounting Software
      • Hosting Services
      • Insurance Solutions
    • Productivity Software
      • Asana vs Trello
      • Cloud Storage Apps
      • Time Tracking Tools
    • Learning Platforms
      • Online Certifications
      • Skillshare vs Udemy
    • Marketing ToolsNew
Reading: Chinese AI Models Capture Majority of OpenRouter Token Volume as MiniMax M2.5 Surges to the Top
Font ResizerAa
WealthariWealthari
  • Home
  • Business
  • Technology
  • Leadership
Search
  • Navigation
    • Home
    • Business
    • Technology
    • Leadership
  • Legal & Info
    • Privacy Policy
    • Terms of Service
    • About Us
    • Contact
© 2024 Wealthari. All Rights Reserved.
Technology

Chinese AI Models Capture Majority of OpenRouter Token Volume as MiniMax M2.5 Surges to the Top

By Michael Scott
Published: February 24, 2026
Share
SHARE

Chinese Models Lead Global LLM Usage

Chinese AI developers have surged ahead in the global language model race, capturing 61% of total token consumption on OpenRouter — the world’s largest large language model (LLM) aggregation platform. According to data released on February 24, 2026, the three most-used models on OpenRouter all originate from Chinese AI labs, underscoring their growing dominance in global AI workloads.

Contents
  • Chinese Models Lead Global LLM Usage
  • MiniMax M2.5 Breaks Records
  • Coding and Agent Workflows Drive the Shift
  • Price Advantage Boosts Adoption

MiniMax M2.5 Breaks Records

At the top of the leaderboard, MiniMax M2.5 posted a record-breaking 2.45 trillion tokens in weekly usage — a 197% increase over the previous week. Data from China’s Science and Technology Innovation Board Daily shows Moonshot AI’s Kimi K2.5 in second place with 1.21 trillion tokens (down 20% week over week), followed by Zhipu AI’s GLM-5 with 780 billion tokens after a 158% surge. DeepSeek V3.2 rounded out the Chinese contingent in fifth position.

Collectively, Chinese models accounted for 5.3 trillion of the 8.7 trillion tokens consumed by OpenRouter’s top ten models. The spike in activity for MiniMax M2.5 was partly driven by temporary free promotions: AI coding platforms like Kilo Code and Cline offered complimentary access to the model in mid-February. Despite the promotional effect, interest in MiniMax M2.5 remains high — the model achieved an 80.2% score on the SWE-Bench Verified benchmark, nearly matching Anthropic’s Claude Opus 4.6 at 80.8%.

Coding and Agent Workflows Drive the Shift

The sharp rise of Chinese models reflects a broader trend in coding and agentic AI usage. Programming tasks now account for more than half of all token consumption on OpenRouter — up from just 11% in early 2025. At the same time, agent workflows, where models autonomously execute multi-step instructions, now generate the majority of output tokens.

OpenRouter COO Chris Clark told SaaStr that Chinese open-weight models are “disproportionately heavy in agentic flows run by U.S. firms.” The three leading Chinese models — MiniMax M2.5, Kimi K2.5, and GLM-5 — are all optimized for coding tasks and automation-driven applications, helping explain their rapid adoption.

Price Advantage Boosts Adoption

Pricing has played an equally critical role in this shift. MiniMax M2.5 costs $0.30 per million input tokens and $1.10 per million output tokens, while GLM-5 lists similar rates. In comparison, Anthropic’s Claude Opus 4.6 charges $5 and $25 per million tokens — making the Chinese alternatives 10–20 times cheaper for developers and startups.

Overall, OpenRouter’s total weekly token usage has climbed to 12.1 trillion, a 12.7-fold increase compared to a year ago. Founded by former OpenSea CTO Alex Atallah, OpenRouter now offers unified API access to more than 400 models from over 60 providers.

According to MIT Technology Review, this cost advantage is fueling widespread adoption of Chinese open-weight models even within U.S. startups. Venture capitalist Martin Casado of Andreessen Horowitz estimates that about 80% of young AI companies using open-source stacks are now running on Chinese-made models.

Disclosure: Wealthari works with brand partners and receives compensation for some recommendations. Our content remains independent and reflects our honest evaluations.
Share This Article
Facebook LinkedIn Email Copy Link

Related

Can a Consultant Benefit Your Business?
What to Know Before Engaging a Business Consultant
June 6, 2025
Why Cybersecurity is Important for Your Wallet
Why Cybersecurity is Important for Your Wallet
June 1, 2025
Why Delegation Matters for Getting the Best Out of Your Team
Why Delegation Matters for Getting the Best Out of Your Team
May 17, 2025

In Business

Nvidia-Led Tech Selloff Drags Wall Street Lower as Block Slashes Half Its Workforce

Shell and METLEN Sign LNG Deal to Bolster Southeast Europe’s Energy Security

Microsoft Unveils Glass Data Storage System That Could Preserve Information for 10,000 Years

  • Privacy Policy
  • Terms of Service
  • Contact Us

Stay updated with insightful content designed to inform and inspire you through the dynamic world of finance, business strategies, and changing market trends.
Hosted by Cloudwork

Wealthari
  • Home
  • Business
  • Technology
  • Leadership
  • Reviews
393.9kFollowersLike
34.3kFollowersFollow
4.42MSubscribersSubscribe
30.4kFollowersFollow

Made by ThemeRuby using the Foxiz theme.