China’s AI sector has a new heavyweight. Beijing-based Moonshot AI has launched Kimi K3, a 2.8-trillion-parameter system that it describes as the world’s largest open-source AI model.
Kimi K3 is designed for long-horizon coding, reasoning, and knowledge work, giving companies another alternative to proprietary models from OpenAI and Anthropic. The launch also adds momentum to China’s open-model ecosystem, although the model’s size could make private deployment impractical for many businesses across Asia-Pacific.
Moonshot’s latest release arrives as Chinese developers compete through faster model launches, lower usage costs, and systems that organizations can modify for their own applications.
Kimi K3 approaches the 3 trillion mark
Reuters reported that Kimi K3 was the first open-weight model to approach the 3 trillion-parameter threshold. Open-weight models allow organizations to download and customize the underlying system, while proprietary models remain controlled by their developers.
Moonshot said the model included a 1-million-token context window, which could help it process large codebases, documents, and extended workflows in a single request.
According to Reuters, Moonshot said Kimi K3 “performed competitively with Fable 5” and substantially outperformed Anthropic’s Opus 4.8, OpenAI’s GPT-5.6 Sol, and GPT-5.5 in GPU kernel optimization tests.
The South China Morning Post noted that Moonshot also acknowledged that Kimi K3 still trailed the strongest proprietary models in overall performance.
Independent evaluations offered additional support for some of the company’s claims. Reuters said that Arena.ai ranked Kimi K3 first in a test focused on building web interfaces, while Vals AI placed it second overall behind Anthropic’s Claude Fable 5.
China’s model race puts pressure on regional rivals
Kimi K3 arrived only weeks after competing releases from Chinese developers such as Z.ai, DeepSeek, and MiniMax. Shares of Z.ai and MiniMax fell sharply in Hong Kong following Moonshot’s announcement, showing how quickly investor expectations are shifting around model performance.
Bank of America analysts said the model demonstrated how Chinese developers could continue improving despite limits on access to advanced computing hardware.
“Despite persistent hardware/compute capacity constraints in China, K3 demonstrates that pre-training scaling, paired with architectural innovation, can still deliver step-change gains for flagship Chinese models,” the analysts said, per CNBC.
The release also adds pressure on other Chinese AI firms to defend their positions in the open-model market.
Lower-cost systems from China are already attracting interest from companies seeking alternatives to leading US platforms, particularly when pricing and customization matter as much as benchmark performance.
More must-read AI coverage
Open access comes with infrastructure tradeoffs
For companies across China and the wider APAC region, Kimi K3 could offer more control over customization, data handling, and local-language applications.
Banks, telecommunications providers, government agencies, and software companies may be especially interested in an open-weight model that gives them greater oversight of how the system is deployed and adapted.
That flexibility, however, comes with a steep infrastructure burden.
Running a 2.8-trillion-parameter model locally could require hundreds of thousands of dollars in computing equipment, according to an estimate cited by Reuters, putting full on-premises deployment beyond the reach of many organizations.
Cloud access may therefore be the more practical route for most businesses. It would reduce the upfront hardware cost, although companies would still need to weigh data residency, vendor dependence, latency, and recurring usage fees.
Size alone also does not guarantee better results.
Businesses evaluating Kimi K3 will need to compare its accuracy, security controls, licensing terms, energy use, and total cost with smaller Chinese models and proprietary US alternatives.
Kimi K3 strengthens China’s position in the global open-model market, but its long-term impact will depend on whether organizations can turn that scale into reliable performance without letting infrastructure costs erase the advantages of open access.
Also read: For a closer look at how lower-cost Chinese AI models compare with OpenAI and Anthropic on pricing, security, and enterprise risk, read TechRepublic’s full analysis.
