Code strings that label “Claude Opus 4.6” and “Claude Opus 4.6 Thinking” have been spotted in what appears to be Perplexity’s model configuration, with identifiers like claude46opus and claude46opusthinking shown alongside a “search” mode.
The discovery suggests Perplexity has already wired in references for a newer Opus variant, at least at the configuration layer, even if it is not selectable for most users yet.
Don’t miss the best of The Mac Observer
Set us as a preferred source and our Apple reporting ranks higher in your Google Search results and Discover feed — one tap, no account changes.
The timing is notable because Perplexity’s own public materials currently point to a smaller, known set of Anthropic options.
Its Help Center lists “Claude 4.5 Opus” as the flagship Claude option for Max subscribers and describes how Perplexity exposes advanced models through its search modes, including toggles for deeper reasoning.
On the API side, Perplexity’s published model list for its Agentic Research API currently includes anthropic/claude-opus-4-5 but does not list an Opus 4.6 entry. And Anthropic’s most recent Opus announcement on its own site is for Claude Opus 4.5, with no public post announcing an Opus 4.6 release.
That gap is why the “spotted” strings should be read as an integration signal, not a launch confirmation. Platforms often stage model IDs ahead of time for testing, partner readiness, internal rollouts, pricing work, or UI toggles that ship before access is granted.
Recent chatter on X has framed the find in exactly that light, as evidence of preparatory plumbing rather than a public release event.
It also leaves room for the most boring outcome: placeholder names that never go live, or that get swapped for a different identifier before launch.
For now, the only reliable confirmation points are still the official Perplexity model roster pages and a formal Anthropic announcement, since those are the places where availability, pricing, and access tiers are made explicit.
Discussion