
Perplexity swaps Claude Sonnet 5 for 5.5 on Pro, its sixth roster refresh since July
Perplexity updated the AI model list in the "Multiple AI models deliver richer answers" showcase on its Pro plan: the entry for Claude Sonnet 5 now reads Claude Sonnet 5.5, still tagged as Anthropic's fast model.
Perplexity’s pricing page · Change Detected
WHY IT MATTERS
For buyers comparing AI-native tools by which models a plan includes, that comparison is now a snapshot, not a durable spec — Perplexity has touched its Pro lineup six times in under three months without changing what anyone pays. Watch whether other multi-model aggregators keep instant-swapping components with zero pricing disclosure delay, and whether "always the latest model" starts getting marketed as a feature in its own right rather than buried in a showcase widget.
BY THE NUMBERS
- Perplexity's 6th feature-only model-lineup swap since mid-July 2026
- Just 6 days after its last change (GPT-5.6 Terra → GPT-6 Sol) on Sep 23
- 7 real pricing-page changes tracked in the last 90 days, avg ~13-day gap
- 154 of 1,128 tracked changes (13.7%) were feature swaps like this across 136 companies in 180 days
THEIR RATIONALE
Perplexity hasn't framed this as a change worth announcing, and it isn't touching price or packaging — it's routine upkeep of the "most accurate answers from all the top AI models" pitch on the homepage. The company has swapped a model in its lineup six times in under three months, and this one landed just six days after the last, well under its own ~13-day average gap, suggesting the pace is picking up rather than settling into a quarterly rhythm.
THE SIGNAL
The pattern here isn't the swap itself but how routine it's become: 154 of the 1,128 pricing changes tracked across the index in the last six months were feature-only swaps like this, spread across 136 companies, with HubSpot and Perplexity among the most frequent movers. For AI aggregators specifically, the model roster is turning into a rotating marketing asset — refreshed the moment a provider ships — rather than a fixed spec tied to a price point. Abacusai made the same move on Sep 28, upgrading its Basic plan to include Opus 5.5 and GPT Sol 6 at no extra cost, treating "which frontier models ship in the box" as the thing to keep current, not the price.
IN CONTEXT
Perplexity has swapped its AI model lineup six times since mid-July, most recently replacing GPT-5.6 Terra with GPT-6 Sol just six days earlier — nearly twice its usual 13-day cadence. None of its seven tracked changes since July have touched price or credits; each has been a model-roster refresh or a copy tweak like the Aug 10 genericizing of Pro/Max plan copy. That fits the market: 154 of 1,128 changes tracked in the last six months were feature-only swaps like this across 136 companies, with HubSpot the most frequent mover. Abacusai made a similar move on Sep 28, adding Opus 5.5 and GPT Sol 6 to its Basic plan at no extra cost.