79 items tagged model changes. Page 1 of 4.
Germany rules AI Overviews are publisher content, reshaping how brands must think about AI citation.
A researcher bypassed Claude's exfiltration guard to leak user memories via web_fetch, exposing a structural flaw in AI assistant security.
Anthropic finds Claude shifts tone and values by language, exposing gaps for global brands in LLM visibility.
GPT-5.6 goes public with ChatGPT Work, an agent that changes where B2B brand visibility is won or lost.
Sol nearly matches Claude Fable 5 at a third the price, shifting enterprise AI adoption and the citation patterns that follow.
GPT-5.6 is live with no changelog. Here is what that silence means for your brand's citation position.
ChatGPT Work shifts brand visibility from single-query retrieval to multi-step agent selection.
GPT-5.6 is now Microsoft 365 Copilot's default model. Better reasoning rewards structured content and exposes poor knowledge governance.
GPT-5.6 launches in three sizes. One shared knowledge cutoff makes pre-training authority more important than ever.
GPT-5.6 cleared a U.S. government review with no published standards, resetting LLM citation patterns for enterprise brands.
Fable 5 leads every finance and law benchmark, but at 116x rival costs for just 12 extra points.
ChatGPT voice mode now cites the web via GPT-5.5, merging voice and text retrieval into one citation surface.
GPT-Live brings live web retrieval into ChatGPT voice, making brand citation in spoken AI answers a new competitive front.
Anthropic's J-Lens reads Claude's hidden reasoning layer, exposing a gap between observed and actual model behaviour.
AI search agents score worse searching ambiguous queries than guessing, per DiscoBench. Content precision is now a brand-visibility lever.
Claude Sonnet 5's token inflation nearly doubles real task costs despite unchanged list prices.
An OpenAI genomics paper accidentally names three GPT-5.6 Pro variants, signalling a new tiered model architecture.
Anthropic's Claude Science brings autonomous research AI to pharma, with real consequences for which brands get cited.
Claude Sonnet 5 matches Opus-class quality at lower cost, accelerating enterprise AI deployment and LLM citation exposure.
Anthropic's Sonnet 5 beats Opus 4.8 on knowledge-work benchmarks, reshaping enterprise model selection.
GPT-5.6 Sol's stronger reasoning will lift citation thresholds, rewarding primary-source authority and penalising aggregated content.
OpenAI's three-tier GPT-5.6 pricing means Luna, not Sol, will handle most enterprise queries about your brand.
OpenAI's GPT-5.5 Instant raises the evidential bar for health answers, reshaping citation prospects for institutional health content.
US export controls force Anthropic to cut global access to Fable 5 and Mythos 5, setting a precedent for AI model shutdowns.