AI Price War Erupts as Cheaper Frontier-Class Models Arrive

OpenAI and Anthropic are turning AI pricing into a full-scale contest, and customers are seeing the impact immediately. OpenAI cut GPT-6 token prices in half, while Anthropic introduced Claude Opus 5.5 at a lower operating cost than Opus 5.

The moves arrive as businesses tighten AI budgets, scrutinize return on investment, and search for ways to control spending. CFOs are reassessing costs tied to AI tools, giving every API price cut a direct path from a company announcement to an enterprise purchasing decision.

OpenAI Pushes GPT-6 Costs Down

OpenAI released GPT-6 Sol and GPT-6 Luna, offshoots of its flagship GPT-6 Astra model. The company also slashed API costs by 50 percent below the pricing on a promotion it is currently running for Astra’s predecessor, GPT-5.6.

That places OpenAI’s release below Anthropic’s pricing, adding another layer to a contest already reshaping the market. Both companies are rolling out cheaper models while announcing outright price cuts on their most expensive models, a combination that puts pressure on customers to try more AI while putting pressure on the companies to earn more from each request.

The cost reductions are tied to improvements in caching and inference. Caching can reduce repeated work, while better inference can help a model produce results with fewer resources. Together, those gains give OpenAI a way to cut prices without changing the story around GPT-6 Astra, GPT-6 Sol, and GPT-6 Luna.

Anthropic Answers With Opus 5.5

Anthropic debuted Claude Opus 5.5, which performs at the level of its flagship Fable 5.1 model. Anthropic claims Opus 5.5 costs around 40 percent less to run than Opus 5, with the company also describing the savings as closer to 40 percent compared with Opus 5.

That creates a different kind of value proposition from OpenAI’s direct token-price reduction. Anthropic is presenting a model that reaches the level of Fable 5.1 while lowering the cost of operation, giving businesses another route to reduce AI spending without stepping away from high-end model performance.

Anthropic also says Opus 5.5 underwent its most rigorous alignment testing to date and was evaluated by outside groups. The testing adds a safety dimension to the price battle, as the companies compete not only on cost but also on how their models perform and how they are assessed.

A Price War Beyond the Frontier

OpenAI and Anthropic are engaged in a price war that is driving down the price of AI and their ability to profit from it. Ara Kharazian, lead economist at Ramp, described the pressure directly: “OpenAI and Anthropic are engaged in a price war that is driving down the price of AI and therefore driving down their ability to profit from it and grow the price of models.”

The releases from both companies are not frontier models, but they show continued progress outside of the frontier. That distinction matters because the competition is no longer limited to a race for the most advanced system. Models that cost less to operate can reach more business workflows, especially when CFOs are measuring every AI tool against its return.

The discussion around the releases also reflects the tension surrounding AI development. One comment captured the contradiction this way: “I’m getting mixed signals — you think these will end the world and you want to pull the ladder up for those pesky open models be regulated, and then you launch new products despite your warnings. I’m confused.”

Another comment offered a warning about the claims surrounding the launches: “The load-bearing word in this sentence is ‘reportedly.’” A separate response joked, “We lose money on every transaction, but we make it up in volume!” Together, the reactions point to the same unresolved question: can lower prices create enough demand to offset thinner margins?

As of September 22, 2026, at 2:00 PM ET, OpenAI and Anthropic have made the direction clear. Caching, inference improvements, cheaper model variants, and lower API prices are pushing AI costs down while businesses demand stronger answers on value. The next phase of the market will test whether that momentum expands AI adoption faster than it erodes the profits behind the models.

Stay connected via Google News

Follow us for the latest AI News