Anthropic drops Fable 5.1 and Mythos 5.1: Coding gods or just cheaper chatbots?
Anthropic just dropped a fresh update for Fable 5.1 and Mythos 5.1, and honestly, it’s a masterclass in making us feel slow. They’re faster, cheaper, and supposedly smarter at everything from coding to dodging digital disasters.
The company overhauled its specialized models to push boundaries in agentic tasks and research. Fable 5.1 scored a massive 52.6% in the Terminal-Bench Science test, doubling the performance of its predecessor. While the token prices remain locked at $10 for input and $50 for output, the cost-cutting hero here is the cache reading, which got a 75% discount, dropping to just $0.25 per million tokens.
Mythos 5.1 enters the ring as the high-stakes sibling, tailored strictly for cybersecurity and life sciences. Both models share the same genetic code, but their behavior is throttled differently by safety guardrails. Fable 5.1 is now the go-to for hunting vulnerabilities in code, while Mythos 5.1 remains locked behind an elite velvet rope for vetted organizations only.
The race to replace developers with glorified autocomplete scripts continues at breakneck speed. While the price drop on caching is a welcome sanity check, the reality of having a silicon brain that is better at finding security holes than an actual human security analyst feels less like progress and more like a ticking clock for anyone writing code for a living.
Source: Anthropic
Comments
This is where the magic happens: AI reads your discussion and rewrites the article based on the most interesting comments. Each strong comment adds points to the meter below. Once the meter is full, the article updates live — no page reload needed.