Key Highlights
- OpenAI cut prices for its GPT-5.6 Luna model by 80 percent, while Anthropic launched its Opus 5 model at half the price of its flagship Fable 5 model
- Prices for models from leading US AI labs have fallen by almost a quarter since mid-July, according to a token price index
- Companies including DoorDash and Airbnb have said they have started using Chinese-made AI models to help control rising bills
- Anthropic this week called off a planned price increase for its Sonnet 5 model that had been due to take effect in September
US AI Labs Cut Prices as Chinese Rivals Gain Enterprise Traction
Leading US AI labs including OpenAI and Anthropic are releasing cheaper models as they compete to retain cost-conscious customers increasingly drawn to lower-priced alternatives from Chinese rivals. The price war comes as rising AI bills prompt companies to curb usage and seek cheaper options, helping Chinese developers including Moonshot and DeepSeek make inroads with users ranging from Silicon Valley to Europe.
OpenAI recently said it was cutting prices for GPT-5.6 Luna, described as its fastest and most affordable model, by 80 percent. Anthropic has launched Opus 5, touting the system's frontier-level intelligence at half the price of Fable 5, the company's most capable model. The moves have helped reduce prices customers pay for models from leading US labs by almost a quarter since mid-July, according to a token price index tracking the units of data processed by language models, which are typically used to calculate customer bills.
Shift Marks Departure From Performance-Only Competition
The cuts mark a notable shift for US AI developers, whose proprietary "closed" models have until now competed primarily on performance rather than price. Increasingly capable "open" Chinese models, which can be freely downloaded and modified by developers, have contributed to growing pressure on pricing across the industry. The moves also come as OpenAI and Anthropic both pursue potential initial public offerings at trillion-dollar valuations, with investors seeking evidence that the industry's substantial AI infrastructure spending can generate meaningful returns.
Corporate customers face additional cost pressure as both companies shift some enterprise clients away from flat subscription pricing toward usage-based billing, under which companies pay according to the computational resources they consume. Some businesses have responded by capping AI usage internally or testing cheaper alternatives, with companies including DoorDash and Airbnb saying they have begun using Chinese-made models to help control rising bills.
Headline Prices Don't Tell the Full Story
OpenAI cut the price of GPT-5.6 Luna from $1 to $0.20 per million input tokens and from $6 to $1.20 per million output tokens. Anthropic launched Opus 5 at $5 per million input tokens and $25 per million output tokens — half the price of its Fable 5 model — and this week called off a planned price increase for its Sonnet 5 model that had been due to take effect from September.
Headline token prices, however, do not provide a straightforward basis for comparison between models, since more capable systems can sometimes complete tasks using fewer tokens or fewer attempts, meaning an apparently pricier model can ultimately cost less for a given task. A benchmarking firm found that Anthropic's Opus 5 running at "medium" effort delivered performance and cost per task comparable to Moonshot's Kimi K3 running at "max" effort, while OpenAI's GPT-5.6 Luna at "max" effort performed similarly to DeepSeek's V4 Flash at "max" but cost nearly twice as much per task. A person close to Anthropic said Opus 5's pricing below Fable 5 reflects how the company's family of models is structured, rather than a response to competitor pricing. A technology lead at a web hosting company that has used large language models since 2020 noted that prices for the very best-performing models remain flat to rising, describing the recent cuts as the first real test of whether firms such as Anthropic and OpenAI can protect pricing on their most advanced offerings.





_11_13_2026_20_30_45_840725.jpg)
Please wait processing your request...