Osegi
AI·2 min read

Anthropic Launches Claude Fable 5.2 and Mythos 5.2, 30% Faster

Anthropic has released Claude Fable 5.2 and Mythos 5.2, promising better coding, reasoning and vision, output up to 30% faster and lower prices than version 5.1.

Abstract illustration of the Claude Fable 5.2 and Mythos 5.2 AI models from Anthropic with code and speed motifs.
Image: اليوم السابع

Anthropic has released Claude Fable 5.2 alongside Claude Mythos 5.2, an update aimed at improving coding, reasoning and image understanding while producing results faster and at lower cost.

According to the company, both models generate output up to 30% faster than Fable 5.1 and Mythos 5.1. Pricing is set at $8 per million input tokens and $40 per million output tokens. Early testing also showed the models communicate more clearly than their predecessors.

Fable 5.2 is the second major update to the Fable 5 tier since its debut in June, following Fable 5.1 in September. Anthropic describes the tier as the most advanced in the Claude lineup, with Haiku, Sonnet and Opus occupying other capability levels.

Who the models are for

Anthropic says Fable and Mythos target complex technical work and large, open-ended projects. For most coding and everyday knowledge tasks, it recommends starting with Claude Opus 5.5 and switching to other models depending on the task's complexity.

On coding, the company says more than 90% of the changes merged into its core codebase are now written by Claude with human assistance. It says it used Fable 5.2 to fix complex bugs within minutes, improve the performance of its services and finish large projects that would otherwise have taken longer or cost more.

One example: a group of AI agents running an earlier version of Fable 5.2 migrated an important production service from Python to Rust. The move improved its efficiency and reliability and made startup about 26 times faster.

Extra usage for subscribers

With the launch, Anthropic is offering additional Fable usage to subscribers on some paid plans this week. Eligible users can reset their Fable usage limits between one and three times, depending on how long they have subscribed. The model will also be available at up to 50% of weekly usage limits for Pro subscribers and some Team plan users.

Apps worth a look

SUMO: the number puzzle

Tap the right tiles, add them up, hit the target. Easy to learn, hard to master. Free, for all ages.

Tags:anthropicclaudefable 5.2mythos 5.2ai modelscoding

More from this day

See all →

More in: AI

See all →