Leading US artificial intelligence laboratories such as OpenAI and Anthropic are releasing cheaper models as they fight to retain cost-conscious customers. These users are switching to cut-price alternatives from Chinese rivals like Moonshot and DeepSeek.
The price war comes as rising bills push companies to curb usage and seek affordable models. OpenAI recently cut prices for GPT-5.6 Luna by 80 percent. The company describes Luna as its fastest and most affordable model. Anthropic launched Claude Opus 5 at half the price of Fable 5, its most capable model. Anthropic also cancelled a planned price increase for its Sonnet 5 model that was scheduled to begin in September.
These reductions have decreased prices paid by customers for leading US models by almost a quarter since mid-July. Silicon Data tracks this metric through its token price index. Tokens measure units of data processed by language models to calculate customer bills.
This pricing shift challenges US groups that historically focused on proprietary closed models competing on performance. Capable open Chinese models that developers can freely download and modify have added to the pressure. The cuts also arrive as OpenAI and Anthropic plan initial public offerings at trillion-dollar valuations. Investors are seeking proof that heavy industry spending on AI can generate financial returns.
Corporate AI users face cost pressures as labs shift enterprise customers away from flat subscriptions toward usage-based billing. Companies now pay according to the computational resources they consume. Some businesses have responded by imposing usage caps or testing cheaper alternatives. Corporations including DoorDash and Airbnb now use Chinese-made models to reduce expenses.
Artificial Analysis benchmarks models across math, science, coding, and reasoning. The firm found Anthropic Opus 5 at medium effort delivered similar performance and cost per task to Moonshot Kimi K3 at max effort. OpenAI GPT-5.6 Luna at max effort performed similarly to DeepSeek V4 Flash at max effort, but cost just under twice as much per task.
A person close to Anthropic stated that Opus 5 pricing was part of the startup model family design rather than a reaction to competitors. Mantas Lukauskas, AI tech lead at Hostinger, noted that prices for the very best models remain flat to rising. He added that recent changes test whether US groups can protect top-tier pricing. US labs cut middle-tier prices while defending top-tier pricing.
