Anthropic has launched Claude Fable 5.2, a new AI model that outperforms Fable 5.1, Opus 5.5, OpenAI’s GPT-6 Astra, and Google’s Gemini 4 Carbon across most benchmarks. The model is 30 percent faster than its predecessor and solved Fermat’s Last Theorem in 30 hours.
On FrontierCode 2.0, Fable 5.2 at High effort matches Opus 5.5 at Max effort for half the cost, while Xhigh effort sets a new record. Pricing is $8 per 1M input and $40 per 1M output tokens. Meta also released Muse Ember, which rivals GPT-6 Luna at lower costs.
The release shifts the competitive landscape, with Fable 5.2 offering High reasoning at $17.94 per test via Vals AI. Anthropic Pro users get a 50 percent weekly usage allowance and free resets. Meanwhile, Muse Ember matches Haiku 5.5 performance for just $0.13.
