Anthropic Releases First Model Since Calling to Pace the Frontier; Pace Turns Out to Be Same Morning as OpenAI
Both labs say the timing was a coincidence, and both labs had the blog post ready at 9 a.m.
SAN FRANCISCO — Anthropic on Tuesday released Claude Opus 5.5, which the company described as "our first release since we called for pacing the frontier," roughly two hours before OpenAI released two new models of its own, GPT-6 Sol and GPT-6 Luna, in what both companies described as an ordinary Tuesday.
"We said the frontier should be paced, and it was," said an Anthropic spokesperson. "It was paced to exactly the speed of the other frontier."
The announcements arrived close enough together that 64 percent of developers surveyed by the analytics firm Cadence Metrics reported reading both launch posts in the same browser window, and 11 percent reported briefly believing they were the same post.
"There is a well-known phenomenon in the industry where two labs happen to finish a year of work on the same day," said Marguerite Osei-Lindqvist, who tracks model releases. "It happens about once a month. Scientists are still trying to determine what causes it, although several of them have noticed that it tends to happen on days when the other lab has a launch."
Opus 5.5 was tested before release by outside evaluators, which the company said demonstrated its commitment to caution. OpenAI said its models were cheaper, which it said demonstrated its commitment to Tuesday.
A person familiar with the matter at one of the labs, who asked which one before agreeing to speak, said the teams do not coordinate. "Nobody calls anybody," the person said. "We simply both check the other company's status page every six minutes until something changes."
Developers said they would evaluate both models carefully over the coming weeks, then use whichever one had not been down that morning.
At press time, a third lab had announced a model at 4:58 p.m. Eastern, describing it as "a quiet release."