Anthropic Weighs New AI Model as OpenAI’s Astra Gains Ground

Anthropic is weighing whether to release a new AI model as OpenAI’s GPT-6 Astra gains commercial momentum. But the reported deliberations are not a launch announcement: the company is still assessing the prospective model’s safety and weighing deployment investment against profitability before a potential IPO.

Reuters, cited by CTech and PYMNTS, reported that Anthropic is considering a release partly in response to Astra. OpenAI released Astra on September 3 and described it as its most capable broadly deployed model. That creates a straightforward competitive rationale for Anthropic to move, while leaving the timing and existence of any release unresolved.

The signal behind the deliberations

The pressure is visible in two usage-and-spending indicators cited by CTech. Ramp’s tracked enterprise AI spending put Astra at about 13%, compared with about 8% for Anthropic’s Claude Fable. OpenRouter also said its users spent more on OpenAI models than Anthropic models in the cited week, its first such lead on that measure in more than two and a half years.

About 13%GPT-6 Astra

Ramp data cited by CTech put Astra at about 13% of tracked enterprise AI spending.

About 8%Claude Fable

The same cited Ramp data put Anthropic’s Claude Fable at about 8%.

A release would have to clear two tests

First is safety. Anthropic is reportedly evaluating the next model’s safety as it considers whether to release it. That is more than a procedural detail: it means a model intended to answer a competitive challenge remains subject to a decision about whether it is ready for deployment.

Second is economics. PYMNTS, also citing Reuters, said Anthropic’s discussions include balancing investment in deploying a new model with efforts to improve profitability as investors focus on when AI companies may produce profits. The potential IPO gives that question extra weight, but it does not turn the reported model consideration into a commitment to ship.

Competition does not settle the decision

OpenAI’s release helps explain why Anthropic is under pressure. Yet the evidence so far supports only an internal choice still being weighed: whether a new model can meet Anthropic’s safety expectations and make commercial sense at the same time. Until that choice is made, the consequential news is the constraint on the race, not another model on the market.