What AI slowdown? OpenAI, Anthropic release dueling models as price wars heat up

Anthropic and OpenAI spent the same day doing the same thing: making their frontier models cheaper. The timing was no accident. Each company has been talking about slowing the pace of AI development, yet both still moved fast enough to release new, lower-cost versions within hours of each other.

Anthropic launched Claude Opus 5.5, saying it matches the performance of its flagship Fable 5.1 while costing about 40% less to run than Opus 5, which shipped in July. It’s also the first model in a new family. Sonnet 5.5 and Haiku 5.5 are expected in the coming weeks, and Anthropic says Opus 5.5 is its most efficient model yet. The company is pairing that with price cuts and higher rate limits.

OpenAI answered with GPT-6 Sol and GPT-6 Luna, both derived from GPT-6 Astra, which arrived earlier this month. The pitch is simple: take Astra’s capabilities and package them for everyday work. OpenAI also cut API pricing by 50% below the promotional rate it’s currently using for GPT-5.6, the model that came before Astra. It says the savings come from improvements in caching and inference, and that it is passing them through to customers.

This is the real story behind the frontier-model theater: the market is moving from “who has the biggest model” to “who can get the better deal.” Ara Kharazian, an economist at Ramp, called it a price war, and that looks right. Businesses are getting pickier, CFOs are asking what the tools actually save, and the old assumption that more capable AI should automatically cost more is starting to look shaky.

Anthropic also said Opus 5.5 went through its most rigorous alignment testing so far and was reviewed by outside groups. It is launching with the same cybersecurity, biology, and frontier-AI-development safeguards previously reserved for Fable 5.1. So even while the companies keep talking about restraint, they’re still pushing the commercial side hard.