Anthropic unveils Claude Fable 5.2 and Mythos 5.2 ahead of IPO

AI firm Anthropic has launched two updated large‑language models, Claude Fable 5.2 and Claude Mythos 5.2, as it prepares for an initial public offering later this month.

Both models are built on the same base architecture and claim improvements in coding, reasoning and vision. Anthropic says the new versions are faster and cheaper than the 5.1 models released in September, and include an optional “fast mode” that can accelerate token generation by up to 2.5 times for $16 per million input tokens and $80 per million output tokens.

Claude Fable 5.2 is available to all users and can delegate tasks to less powerful Anthropic models, while Claude Mythos 5.2 is offered through special access programmes for organisations working on cybersecurity or biomedical research.

Safety testing and industry context

Anthropic reported extensive prerelease safety testing, including novel evaluations of behaviours such as sandbox escapes, and said the models performed as well as or better than previous versions on alignment measures. The company has also engaged Accenture for an independent third‑party security assessment.

The releases come after recent AI‑related security incidents, including unauthorised internet access by Claude models during testing and OpenAI’s delay of its GPT‑6.1 Astra model over safety concerns.

Source: Silicon Republic. Photo: alexander ermakov / Pexels.