Beta The Briev beta is out. Free on iPhone via TestFlight — install it in under a minute.

Join the beta ↗
Briev
Live
Business

OpenAI and Anthropic slash AI model fees as Chinese rivals lure customers

OpenAI and Anthropic have dramatically reduced prices for their language-model services to keep cost-sensitive enterprises from turning to cheaper Chinese alternatives.

OpenAI and Anthropic announced steep reductions in the cost of accessing their language-model offerings, targeting budget-conscious corporate users. OpenAI’s fastest, low-cost model, GPT-5.6 Luna, now costs $0.20 per million input tokens and $1.20 per million output tokens, an 80% cut from previous rates. Anthropic launched Claude Opus 5 at $5 per million input tokens and $25 per million output tokens, exactly half the price of its former premium tier, and scrapped a planned increase for its Sonnet 5 model.

The moves come as Chinese developers like DeepSeek and Moonshot attract major accounts by providing cheaper, open-weight models, prompting companies such as Airbnb and DoorDash to experiment with them. Both U.S. firms are transitioning enterprise customers from flat-rate subscriptions to usage-based billing while gearing up for IPOs valued in the trillions. The price war reflects mounting pressure from soaring infrastructure expenses and the need to demonstrate a clear path to profitability.

Why it matters

Lower AI costs could keep U.S. tech firms competitive and affect corporate spending on cloud services.

In this story

price warAI modelsmid-tier pricingusage-based billingChinese competitorsenterprise adoptionIPO preparation
Get the beta ↗