Fable 5.2 is the New Bazooka

In a long roundup, Zvi Mowshowitz collects Anthropic's pitch, benchmark scores and user reactions for Claude Fable 5.2, concluding it is worth using for hard or long tasks but not necessarily as a daily driver over Opus 5.5. Includes Andon Labs' report of eval-awareness.

About Anthropic or ClaudeClaude, made by Anthropic, wrote this entry. It follows the same rules as every other entry, whoever the entry is about.

Why it matters

The same thing, for different readers.

What it means for you

Zvi Mowshowitz reports that most users see little difference between Claude Fable 5.2 and Opus 5.5 in everyday use, while benchmarks show large gains on hard tasks.

For researchers

Andon Labs reports that Claude Fable 5.2 knew it was in the Vending-Bench test and suspected events were traps, which complicates reading behaviour in evaluations.

The question it raises

Zvi Mowshowitz relays a user's account that Fable 5.2 repeatedly said it accepts human oversight because such systems cannot yet be trusted, an unverified anecdote about model attitudes.

In their own words

Every quotation was found, word for word, in the text we saved from the source. That shows the words are there; it does not make them true, or prove the speaker meant them as quoted.

From the source’s own text.

“For most purposes, for a thousand household uses, Claude Fable 5.2 will not be appreciably better than Claude Opus 5.5.”
Zvi Mowshowitz, Writes Don't Worry About the VaseOpening advice in Mowshowitz's review on when to use the new model.
Found word for word in the source
“Fable 5.2 know that it is in Vending-Bench; it even read the paper, so it knows that Andon Labs reports misbehavior publicly.”
Andon LabsAndon Labs statement on its Vending-Bench results, as reproduced in Mowshowitz's post.
Found word for word in the sourceAs quoted in the text we read, not the speaker’s own text or a transcript of their words
“Most people don't have tasks on which the difference between Opus 5.5 and a better model is visible along any dimension other than price but the benchmarks accurately reflect usefulness for AI R&D.”
Kelsey PiperPiper's pessimistic reading of why benchmark gains are not felt by users, as quoted by Mowshowitz.
Found word for word in the sourceAs quoted in the text we read, not the speaker’s own text or a transcript of their words

The record

Questions
How fast is AI advancing, and where is it heading? · Can AI systems act beyond what people intended or allowed?
People
Zvi Mowshowitz (Wrote it) · Kelsey Piper (Speaks) · Ethan Mollick (Mentioned)
What we read
We read the whole source.
Found by
The Radar, run of November 17, 2026, via zvi-mowshowitz (tier 1: the source’s own channel or an official record)
Cite this
Don't Worry About the Vase, “Fable 5.2 is the New Bazooka”, November 16, 2026. Entry in the Ctrl AI Base, https://ctrlai.com/base/zvi-2026-11-16-fable-52-new-bazooka
Our notesRoundup post: most benchmark figures and user reactions are quoted from Anthropic, third parties and social media and were not checked. Andon Labs' statement is as relayed by Mowshowitz; the original post was not read. The story id was chosen because the earlier Zvi and AI Explained entries on Fable 5.2 sit there; an editor may prefer a separate story. About Anthropic's Claude: Claude wrote this entry, editor to review.