scripod.com
Claude Fable 5 review: what the new Mythos model gets right (and very wrong)

Overview

Shownote

Highlights

Transcript

Chapters

Pins

Claude Fable 5 review: what the new Mythos model gets right (and very wrong)

How I AI

11 HOURS AGO
How I AI

How I AI

11 HOURS AGO
OverviewShownoteHighlightsTranscriptChaptersPins

Shownote

Claude Fable 5 is the first Mythos-class intelligence model to be generally available, and I got early access to test it before launch. In this episode, I walk through what Anthropic is promising, what actually stood out when I used it on real work, and wh...

Highlights

Claire Vo shares her early access impressions of Anthropic's new Claude Fable 5 model, the first Mythos-class intelligence model to be generally available. She walks through what the model promises, how it performed on real-world tasks, and where it fits in your AI stack.
00:00
Claude Fable 5 (Mythos) early access impressions
00:32
Claude Fable 5 is state-of-the-art, expensive, and excels at benchmarks.
05:14
Token intensity does not guarantee correct results
06:30
95% of sessions did not hit a fallback
07:47
Fable includes safeguards and is generally available.
08:30
Fable 5 as a senior advisor with cheaper execution models
09:22
Claude Fable 5 significantly outperforms other models
09:56
Claude Fable 5 excels at vision and formatting.
11:43
Output detailed but nearly impossible to parse
12:56
Claude Fable 5 produced terrible design results
14:05
Execution too conservative, producing minimal MVPs.
14:43
Multi-agent orchestration had both successes and failures
15:40
Excels at hard technical problems and vision tasks

Chapters

Introduction: Fable 5 is finally here
00:00
What Anthropic says about the model
00:31
Token-intensive by design
05:14
Safety classifiers and the new fallback concept
06:28
Is this or is this not Mythos?
07:46
New product launches: Managed Agents and more
08:30
Crushing benchmarks
09:20
What it’s actually like to use (the good and the bad)
09:55
Test 1: product graph spec
11:40
Test 2: designing a skills registry
12:56
Conservative on execution
14:04
Test 3: multi-agent orchestration
14:43
My takeaways
15:39

Transcript

Claire Vo: It's here! The model, the myth, the legend, Mythos from Anthropic has finally dropped. Well, baby Mythos. We're calling it Fable 5, and this new model is crushing benchmarks, but the question is, can it crush my backlog? I got early access to th...