Gemini

Batt

Recall

A tech columnist says <a href="https://yellow.com/news/claude-opus-4-7-mythos-safety" rel="noopener noreferrer nofollow">Anthropic's Claude</a> handles long-term memory better than <a href="https://yellow.com/news/gemini-flash-2-points-claude-opus" rel="noopener noreferrer nofollow">Google's Gemini</a> because users can read and edit what it remembers.
<h2 id="claude-memory-vs-gemini-recall">Claude Memory vs. Gemini Recall</h2>
Writer Simon Batt<a href="https://www.xda-developers.com/claude-memory-beats-gemini/" rel="noopener noreferrer nofollow">argued</a> at XDA that Anthropic's Claude manages memory better than Google's Gemini.
Claude keeps a central memory log, a block of text describing what the assistant has learned from past conversations. Users can read it. They can also create separate memory files for different projects, keeping personal and professional use apart.
Gemini works differently. Batt said the assistant builds its picture of a user by drawing from every past chat, with no master file to inspect or trim.
<blockquote>
 To make Gemini forget something, he said, a user appears to have to delete the entire conversation that holds it.
</blockquote>
Batt also said Gemini surfaced details that had nothing to do with the topic at hand. While discussing local property prices, he wrote, the assistant paused to ask whether the question related to a science lab he had visited four months earlier.
Also Read:<a href="https://yellow.com/news/goldman-walks-away-xrp-solana-q1-reset" rel="noopener noreferrer nofollow">Goldman Sachs Walks Away From XRP, Solana In Sharp Q1 Crypto Reset</a>
<h2 id="why-memory-control-matters">Why Memory Control Matters</h2>
The complaint points to a wider design split among AI assistants.
Batt's core objection was that Gemini gave him a single switch, on or off, with no setting to adjust what the assistant retained. Anthropic's approach, by contrast, lets a user <a href="https://lumichats.com/blog/chatgpt-memory-vs-claude-memory-vs-gemini-personal-intelligence-2026-which-ai-actually-knows-you" rel="noopener noreferrer nofollow">edit</a> Claude to forget or change specific details without scrapping any chats.
He also flagged a quality issue.
<blockquote>
 Because Claude's memory reads more like a summary than a full transcript, he said, the assistant avoids stray, months-old facts that no longer apply.
</blockquote>
Batt did acknowledge Google's logic. Drawing context from old chats spares users from repeating themselves, he wrote, but treating raw conversations as the memory itself drags irrelevant data into new sessions.
Anthropic <a href="https://www.knightli.com/en/2026/05/07/chatgpt-claude-code-gemini-memory-comparison/" rel="noopener noreferrer nofollow">rolled</a> out automatic cross-conversation memory across all Claude plans, including the free tier, on Mar. 2, with controls to view, edit and delete stored entries in settings. Google has been steadily widening Gemini's reach into user data, drawing scrutiny from reviewers over how much personal information the assistant absorbs and how little of it people can see.
Read Next:<a href="https://yellow.com/news/bitcoin-demand-4-month-low-consolidation" rel="noopener noreferrer nofollow">Bitcoin Demand Crashes To 4-Month Low, Risks Deep Consolidation Phase</a>

Claude Beats Gemini Because Of One Setting You Can Actually Touch, Expert Says

Table of Contents

Yellow News

Other articles published on May 22, 2026

Join our free newsletter for daily crypto updates!