Anthropic lifts the veil on the latest features for Fable 5.2 & Mythos 5.2
Fable and Mythos 5.2 are faster, more accurate, and significantly cheaper than previous versions.
Anthropic has unveiled Fable 5.2 and Mythos 5.2, two new models designed to tackle complex problems and ambitious projects. Both models boast a 20% cheaper token cost compared to their predecessors. Renowned for their prowess in coding, biology, scientific, and mathematical research, Fable 5.2 and Mythos 5.2 deliver more accurate outputs at a lower cost than previous versions.
Fable 5.2 is accessible to all users and boasts enhanced inference capabilities alongside improved cybersecurity and biology safeguards. To access the expanded features of Mythos 5.2, users must apply through Anthropic's Life Sciences and Cyber Verification Programs. Anthropic emphasizes that these models are ideal for "ambitious, open-ended technical work" rather than daily use, recommending Opus 5.5 for such purposes.
The latest Fable release is 20% cheaper than its predecessor. Fable 5.2 costs $8 per million input tokens and $40 per million output tokens. For those with tight deadlines or needing quick results, Anthropic offers a "fast mode" that boosts token generation speed by up to 2.5 times, but at a higher cost of $16 per million input tokens and $80 per million output tokens. Across a range of benchmark tasks, Fable 5.2 outperforms Opus 5.5, providing quicker, more accurate results with a significantly reduced token cost.
Anthropic's engineers demonstrated Fable 5.2's capabilities by having it design a prototype operating system over eight days, involving over 1,000 agents writing, testing, and reviewing code. The result was a fully functional operating system capable of booting and incorporating built-in Claude functionality. "Claude Fable 5.2 is excellent at taking a loose idea and refining it," said Tom Pritchard, Staff Engineer at Shopify.
"Its answers are well-considered, and within a few turns, I can almost always say 'go do that' and trust it to complete the task with minimal intervention. It feels like working with another staff-level engineer." Anthropic showcased Fable 5.2's capabilities across various scientific and mathematical tasks, including formalizing the proof of Fermat's Last Theorem in 30 hours, creating a home voice-assistant hub, a USB speakerphone, and a handheld virtual pet, and solving two open problems in black hole physics.
Written by urgent.news from TechRadar's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
