Mozilla’s latest State of Open Source AI report finds Chinese open-weights models nearly matching US closed frontier systems, with only a 4.4‑month performance gap while costing as little as one‑fifth per task. Benchmarks from Vals AI and METR show models like Moonshot’s Kimi K3 and Zhipu’s GLM 5.2 closely trailing Anthropic’s frontier models, leading many organizations to default to cheaper open options for routine workloads. The report warns about geographic concentration of AI power, urges US and European labs to invest in open ecosystems, and calls for publicly funded compute, neutral foundations, and philanthropic support for audits and evaluations.
This update represents a notable development in the Ai sector. Organizations and founders tracking this space should evaluate potential strategic and technical implications on their operations.