OpenAI and Anthropic Slash Prices to Combat Chinese Rivals
OpenAI and Anthropic reduced model pricing and promoted a new cost-per-task metric to retain enterprise customers switching to cheaper Chinese AI alternatives.
AI developers OpenAI and Anthropic have entered a price war to counter the growing influence of Chinese competitors like DeepSeek and Moonshot. OpenAI reduced the price of its GPT-5.6 Luna model by 80 percent, while Anthropic launched Claude Opus 5 at half the cost of its Fable 5 model. These cuts follow a trend of enterprise customers, including DoorDash and Airbnb, adopting lower-cost open-source models to reduce operational expenses.
Alongside these price reductions, the two firms are promoting a new industry metric called cost per task to replace the traditional cost per token. They argue that token-based pricing is an imprecise proxy for business expenses because cheaper models often require more attempts or human review to succeed. In contrast, more capable models can often complete the same task in a single pass, potentially lowering the total cost of completion.
The shift occurs as both companies face investor pressure to demonstrate returns on massive spending ahead of planned initial public offerings at trillion-dollar valuations. While GPT-5.6 Luna is now considered affordable under the task-based metric, Anthropic's Claude Fable 5 remains among the most expensive options available.
