Claim: Anthropic and OpenAI both hold models far beyond Opus 5.5

ChrisGPT · x · 2026-09-25

An X user claims Anthropic is "sandbagging" — and that both Anthropic and OpenAI possess models significantly more powerful than the released Opus 5.5 and Astra. The assertion is unverified, offers no evidence, and has drawn no official response, but reflects ongoing community suspicion that frontier labs are holding back their strongest models.

Related event: Insiders claim OpenAI and Anthropic hold models far ahead of public releases(2 posts)→

Original post →

More from Models

Models channel →