What happened
TechCrunch's Russell Brandom argued that the familiar OpenAI-versus-Anthropic framing no longer captures the moment. With GPT-5.6 entering the same government-gated preview that has held Anthropic's Mythos for months, both labs now share the same constraint rather than competing on open release.
The Information first reported that GPT-5.6 would ship only in a limited preview approved customer by customer. Anthropic's Mythos had already been in preview for months with no general release, and both cases turn on concerns about cybersecurity capabilities and biorisk and alignment.
Why it matters
The economic risk is shared too: if releases stall, both companies face the same drag on revenue and momentum. That reframes the rivalry as a common regulatory problem rather than a head-to-head race.
By August, an executive order requires the administration to establish a classified process to assess covered frontier models, which would formalize the approval bottleneck the labs are already navigating.
MintedBrain take
The competitive narrative matters less than the release calendar. Watch when and how the classified assessment process lands, because that timeline now governs when the strongest models actually reach builders.
Discussion
Sign in to comment. Your account must be at least 1 day old.