Anything Engine Optimization The rolling record of the AI-search industry.
A minimalist winter scene of a snowy island in Kemi, Lapland, Finland.
Photo by Markku Soini on Pexels
Profound study — vendor-published
Profound research blog · Oct 13 ResearchAEO

Profound: Sonnet 5.5 asks for context 5x more than Sonnet 5

What this answers

Does Claude Sonnet 5.5 ask more clarifying questions than Sonnet 5?
Yes — per Profound's analysis of 28,440 answers, Sonnet 5.5 asked readers for context in 60.6% of answers, versus 12.5% for Sonnet 5.
Does Claude Sonnet 5.5 cite fresher web pages than Sonnet 5?
Yes — Profound found Sonnet 5.5's most-cited pages have a median age of 107 days, versus 164 days for Sonnet 5, with its top 1% of cited pages at 76 days.

Profound ran 711 prompts through Anthropic’s API 20 times each on Claude Sonnet 5 and Sonnet 5.5 with web search enabled — 28,440 answers total — and found Sonnet 5.5 asked readers for context (details about their specific situation) in 60.6% of answers, against 12.5% for Sonnet 5. Sonnet 5.5 also named the current year in 59% of its query fan-outs, more than double Sonnet 5’s rate, and its most-cited pages skewed newer, with a median age of 107 days against 164 days for Sonnet 5. Per Profound, both models carried roughly 12 inline citations per searched answer.

Why it matters: The shift toward more context-seeking and fresher sources suggests AEO content that states dates clearly and speaks to a reader's specific situation is more likely to be used by newer models.

The record: AnthropicProfound

Glossary: Query fan-out

Via Profound research blog ↗

Posted to the wire October 13, 2026. Edited by Joe Balewski.