Claude Fable 5.2 tops AI leaderboard with price cut
Anthropic’s Fable 5.2 is cheaper, faster, and still at the top of the leaderboard.
What Fable 5.2 actually changes
Why pacing, not pausing, matters
The race is still moving fast
Anthropic has released Claude Fable 5.2 and Mythos 5.2, and the pair immediately claimed the No. 1 spot on Artificial Analysis' Intelligence Index. It's the first concrete output of CEO Dario Amodei's call to pace the frontier, and it shows that even a deliberately slower release cadence can outrun the competition.
## What Fable 5.2 actually changes Fable 5.2 is 20% cheaper and up to 30% faster than Fable 5.1, with pricing at $8 per million input tokens and $40 per million output tokens. That still costs about twice as much as Opus 5.5, so this is not a budget-tier model. Yet Anthropic says a team of 1,000 Fable agents built an operating system in eight days and formalized the proof of a famous math problem in 30 hours. The twist: Fable is deliberately kept near Opus level in certain development-related skills, leaving headroom for future upgrades.
## Why pacing, not pausing, matters Anthropic executive Dianne Penn describes the approach as pacing, not pausing. For developers, that means the roadmap emphasizes stable, safety-conscious improvements over rushing to ship every internal breakthrough. The trade-off is that some capabilities stay gated: Mythos 5.2, the same base model with lighter guardrails, is limited to vetted security and life-sciences groups, while Fable ships with stronger cyber and biology protections.
## The race is still moving fast Google's Gemini 4 Carbon hasn't even been released to the public, and Fable 5.2 already edged it out by five points on the Intelligence Index. Meanwhile, Meta's Muse Ember undercuts the cost side, matching Haiku 5.5 at 36 points while running a typical test task for about $0.14 versus $0.41. For teams picking models today, the lesson is that the frontier is no longer just about raw IQ scores—capability, price, and guardrails all factor into the decision. Anthropic's pacing may look slow from the inside, but from a developer's perspective it still feels fast enough to lead.
Test Fable 5.2 if you want frontier-level performance with deliberate safety guardrails—just budget for the premium price.