Coinbase is adopting Chinese AI models like GLM 5.2 and Kimi 2.7 to reduce expenses. The shift demonstrates growing pressure on Western AI labs to compete on pricing.
CEO Brian Armstrong announced the company switched to Chinese alternatives after implementing an automated routing system that selects the most cost-effective model for each task.
The move delivered significant savings. Coinbase halved its AI spending while token usage continued climbing, suggesting improved efficiency rather than reduced usage.
A key optimization came from better caching, which boosted the hit rate from 5 to 60 percent. This means the system can reuse previous results more effectively, reducing redundant processing costs.
The decision reflects a broader trend in tech. As Western AI providers like OpenAI and Anthropic maintain premium pricing, companies face mounting pressure to explore alternatives. Chinese models, increasingly competitive in capability, offer lower-cost options.
Coinbase's approach—using an intelligent router rather than committing to a single provider—allows flexibility to switch between vendors based on task requirements and pricing changes. The strategy highlights how the AI market is fragmenting beyond dominant Western players.
Runway has introduced Solaris, an AI system that renders software interfaces frame by frame as users interact with it, rather than executing traditional code. The model marks the first entry in what Runway calls the "Interface World Models" category.
A researcher achieved 44% accuracy on the ARC-AGI-1 benchmark with a cost of just 67 cents per evaluation. The result highlights the feasibility of solving the abstract reasoning challenge with minimal computational expense.
Artificial intelligence is expanding its impact across white-collar and service sectors, with PR professionals, web developers, and hospitality workers now facing displacement risks as automation capabilities broaden.