AI Price War Heats Up: OpenAI Cuts Enterprise Rates by Up to 80%

Major AI companies that have been competing on AI model performance are now diving headfirst into a price-cutting war.

OpenAI announced on the 30th local time that it will slash enterprise usage fees for the mid- and lower-tier versions of its latest AI model, GPT-5.6, unveiled just three weeks ago, by up to 80%.

The fee for the mid-tier model, GPT-5.6 Terra, will be reduced by 20% to between $2 and $12 per million tokens, while the price for the budget model, GPT-5.6 Luna, will be slashed by 80% to between $0.2 and $1.2 per million tokens.

This price reduction applies to application programming interfaces (APIs) primarily used by developers, with no changes to individual ChatGPT subscription fees.

However, usage limits for these models on personal ChatGPT accounts will be increased in proportion to the price cuts.

While the price of the top-tier model, GPT-5.6 Sol, remains unchanged, a new mode has been introduced that offers 2.5 times faster processing speeds at double the price.

OpenAI explained that this price cut is the result of overall infrastructure efficiency improvements, including enhancements to models, inference systems, and agent connection methods.

Industry observers interpret OpenAI's move to lower prices for its latest models as a strategic effort to boost revenue by expanding its share in the enterprise market, which has been staked out by its rival Anthropic ahead of an anticipated initial public offering (IPO).

This is because competition in mid- to low-priced models has become increasingly crucial as companies, which once drove up AI usage in a trend known as token maxing, have recently expressed frustration over soaring AI costs.

Anthropic also launched Claude Opus 5 on the 25th—a model comparable to its top-tier Claude Fable 5—pricing it at half the level of Fable 5 to attract users.

In addition, the rise of Chinese open-source models, including Moonshot AI's Kimi K3, is evaluated to have influenced the price-cutting competition among U.S. AI companies.
※ Please note: This article was translated by AI and may contain errors.