Anthropic's Fable adds strict guardrails, draws ire
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Anthropic released Fable, a public, limited version of its Mythos cybersecurity AI, on Tuesday.
- Cybersecurity researchers say the model’s guardrails reject innocuous tasks such as reading blog posts or code reviews, deeming them too strict.
- Developers on social media report that the guardrails block any request even tangentially related to cybersecurity, biology, or chemistry, pushing them toward alternative models like Opus 4.8.
- Pricing for Claude Fable 5 and Mythos 5 is set at $10 per 1 M input tokens and $50 per 1 M output tokens, less than half the price of the Mythos Preview.
- Wall Street Journal noted that Anthropic’s Fable is a Mythos‑class model released to the general public with built‑in guardrails.
Why it matters: Developers and security researchers lose productivity because Fable’s guardrails block routine tasks like reading blogs or reviewing code, while Anthropic gains a safety‑first narrative that attracts enterprise customers seeking strict cyber safeguards but risks limiting broader adoption and driving users to competing models.
