The release of Kimi-K3 has reduced the time lag between the open-source and closed-source AI frontiers to just 1.5 months, according to Artificial Analysis.
- Kimi-K3 is still behind models like Fable and Anthropic's offerings but is closing the gap rapidly.
- The model features approximately 3 trillion parameters, indicating that scaling laws continue to drive performance improvements.
- Its size makes it unsuitable for consumer hardware like MacBooks.
- Google's position in the frontier landscape remains unclear following the November release of Gemini 3 Pro.