GoKawiilMozilla report: Chinese open-weights AI models now trail US frontier models by just 4.4 months
A new State of Open Source AI report from Mozilla, shared with Ars ahead of its September 15 release, finds that top open-weights models from Chinese developers now perform nearly on par with closed frontier models from US firms like Anthropic and OpenAI. It cites Moonshot AI's Kimi K3 scoring only three points below Anthropic's Fable 5 on the Artificial Analysis Intelligence Index, despite costing roughly 30 percent as much to run.
GoKawiil's interpretation of the reporting above, not reported fact.
As the performance gap shrinks, many organizations can shift routine AI workloads to cheaper open models and reserve costly closed systems for specialized tasks like expert reasoning, heavy retrieval, or long-context work. This shift could pressure US AI giants' pricing power while boosting the influence of Chinese AI labs in enterprise adoption, even though open models still require more in-house technical expertise to deploy effectively.
- The performance gap between top open Chinese models and closed US frontier models has narrowed to 4.4 months.
- Moonshot AI's Kimi K3 nearly matches Anthropic's Fable 5 in benchmark scores at about 30 percent of the cost.
- Closed models still justify their premium for specialized workloads requiring compliance support and deep expertise.
Source: arstechnica.com, 2026-09-15
Published there as: “Exclusive: Open Chinese models close gap with Silicon Valley’s frontier AI models”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.