Anthropic Told AI Labs to Slow Down; Now It’s Weighing Another Model as OpenAI Pulls Ahead

A week after Anthropic CEO Dario Amodei called on the AI industry to slow the pace of capability development, his company is reportedly considering a new model to respond to OpenAI’s recent momentum.

Reuters reports that Anthropic is considering releasing a new model “to counter OpenAI’s recent launch of GPT-6 Astra,” with three people familiar with the matter saying the company is weighing the move. The report says there is no final decision on whether to release the model or when it might launch.

The timing creates a tension inside a company that has made AI safety central to its public identity. It does not establish that Amodei has abandoned his position. Reuters reports that Anthropic is weighing safety considerations as part of its deliberations. What it does show is the pressure created when a competitor begins gaining ground quickly.

The Numbers Behind It

GPT-6 Astra appears to have gained traction with enterprise customers since its launch. Reuters cited data from corporate-spending platform Ramp showing Astra at roughly 13% of tracked enterprise AI spending, compared with about 8% for Anthropic’s Claude Fable.

Astra has also reportedly risen to the top of OpenRouter, a marketplace used to access different AI models. Those figures do not mean OpenAI has overtaken Anthropic across the entire AI market.

Reuters puts Anthropic’s annualized revenue run rate at about $65 billion, compared with roughly $40 billion for OpenAI. Enterprise spending tracked by one platform and company-wide revenue are different measurements, so the Astra figures show competitive momentum rather than a wholesale reversal of Anthropic’s position.

There is also pressure from an important customer. Meta, historically one of Anthropic’s largest customers, is reportedly looking to reduce its reliance on Anthropic’s models as it develops more AI capability internally, according to people familiar with the matter cited by Reuters.

Amodei’s Slowdown Argument

Anthropic’s own measurements initiative says: “As the world considers slowing the pace of frontier AI development, the public needs more information.” The company says it wants to measure how much AI is involved in building the next generation of AI, how well AI-agent actions can be overseen, and how computing resources are allocated.

OpenAI CEO Sam Altman and SpaceX CEO Elon Musk publicly supported the call. Amodei’s position was not simply about releasing fewer products. It was an argument that the speed of capability development itself is becoming a safety problem.

“We plan to embed independent third-party evaluators from multiple organizations at Anthropic, and give them access to internal processes, systems, and data comparable to what internal risk assessment teams have,” Anthropic says. The company says the evaluators will verify safety practices, report incidents and monitor development metrics.

The IPO Pressure

Reuters reports that Anthropic is considering the model release “ahead of a potential IPO.” The company is weighing the competitive pressure from OpenAI alongside its position with enterprise customers and its plans for the public markets.

Reuters puts Anthropic’s annualized revenue run rate at about $65 billion as of July, versus roughly $40 billion for OpenAI. That creates a second clock for Anthropic.

Model development requires enormous amounts of computing power and capital. At the same time, delaying a new model while a rival gains enterprise traction carries its own commercial risk. Anthropic therefore has to weigh the cost of building and releasing another model against the cost of allowing OpenAI to take more of the market momentum.

Anthropic says its new measurements are intended to give “the public, third parties, and governments better visibility into the pace of AI development inside frontier AI labs.” The company says the metrics cover AI’s role in AI R&D, oversight of agent actions and compute allocation.That makes the next model decision unusually revealing.

If Anthropic releases a new system soon, it will show how competitive pressure can compress the timetable even at a lab publicly arguing for restraint. If it delays, that would demonstrate that its stated safety concerns can outweigh the immediate pressure created by a rival’s gains.