Elon Musk admits, 'I was wrong,' says Anthropic is the AI leader and calls Mythos the best model ever

Elon Musk has made a surprising admission in the rapidly evolving artificial intelligence race. In a post on X, the SpaceXAI founder acknowledged that he had underestimated Anthropic, saying the company is now "obviously currently the leader in AI."

The remark marks a sharp reversal from September last year, when Musk wrote that "winning was never in the set of possible outcomes for Anthropic."

Responding to a user who reposted his earlier prediction, Musk said: "I was clearly wrong about Anthropic. They are obviously currently the leader in AI." He also praised Anthropic's latest Claude Mythos and Claude Fable models, calling them the strongest AI systems released so far.

Musk says no rival has matched Mythos

Musk did not stop at recognising Anthropic's progress.

He wrote that "No company has released a model as good as Mythos/Fable and they will undoubtedly have Mythos 2 ready soon."

The comments are notable because they come shortly after SpaceXAI launched Grok 4.5, while OpenAI recently released GPT-5.6. Both companies have been competing aggressively with Anthropic in coding, reasoning and enterprise AI. Earlier this year, Anthropic's Mythos and Fable models faced temporary restrictions from US authorities over cybersecurity concerns before later returning with updated safeguards.

Rivals in AI, partners in computing

The relationship between Musk and Anthropic has become increasingly complex. Although the companies compete directly in artificial intelligence, Anthropic is also one of the largest customers of SpaceXAI's Colossus 1 supercomputer. According to previously reported details, Anthropic agreed to lease computing capacity worth about $1.25 billion per month through May 2029, making the overall agreement worth roughly $40 billion. The Colossus system reportedly contains more than 220,000 Nvidia GPUs with approximately 300 megawatts of computing capacity.

The arrangement has prompted speculation over whether Musk could use that leverage against a competitor.

"That's not my style"

Musk dismissed those suggestions. When one X user claimed he could effectively shut down Anthropic by ending the computing agreement, Musk replied that he would never cut them off in a way that hurt them badly, even though they are competitors. He pointed to Tesla's open-source patents and its decision to make the Supercharger network available to rival carmakers as examples of his approach to competition.

The episode also highlights another growing issue in the AI industry model distillation, where companies study rival AI systems to improve their own models. During his legal dispute with OpenAI, Musk acknowledged that AI companies generally learn from one another's models, while Anthropic has previously accused some Chinese AI developers of distilling Claude. Musk's latest comments suggest that despite fierce competition, he believes Anthropic currently sets the benchmark in frontier AI and that recognition has become difficult even for rivals to ignore.