Why it matters
If you pay for top-model work, this is a comparison worth looking at: Anthropic says Fable 5.2 gets about level with Opus 5.5 for roughly half the cost per task.
Will your bill really halve? Per task, that's what Anthropic says. I think the top model now has to earn its price.
What came out
What came out

Claude Fable 5.2 launched on 9 November 2026 for everyone, on all platforms.
According to Anthropic, it scores about level with Opus 5.5 at its strongest setting. It also has a twin, Claude Mythos 5.2, with the same abilities, available only to organisations Anthropic has verified.
Science makes a big jump
Science makes a big jump

Anthropic reports a science research test score of 74.6%, up from 52.6% for the previous Fable.
Anthropic also says Fable 5.2 solved two open problems in the physics of black holes. These are the company's reported results.
Roughly half the cost per task
Roughly half the cost per task

Anthropic's comparison is cost per task: Fable 5.2 scores about level with Opus 5.5 at its strongest setting for roughly half the cost. That is the company's claim, rather than a measured saving on your workload.
| Comparison | Anthropic's reported result |
|---|---|
| FrontierCode 2.0 score | 45.3% for Fable 5.2 at High effort, 45.5% for Opus 5.5 at Max effort |
| Cost per task | Roughly half for Fable 5.2 |
Scores and relative cost are Anthropic's own reported figures.
Who gets it
Who gets it

Fable 5.2 is available to everyone, on all platforms. It is included in the Pro plan and standard Team seats, for up to half of the weekly limits.
The verified-organisation restriction applies to its twin, Claude Mythos 5.2.
Read the sources
Read the sources
The launch details, performance comparison and access terms come from Anthropic's announcement.
- Read the announcementClaude Fable and Mythos 5.2
Read Anthropic's announcement for its reported scores, cost-per-task comparison and availability details.
Yes, that's my AI avatar in the video. Same opinions, better lighting.
