Anthropic releases Claude Fable 5 and Claude Mythos 5

Anthropic has launched Claude Fable 5, described as a Mythos-class model made safe for general use. The company says its capabilities exceed those of any model it has previously made generally available and that it is state-of-the-art on nearly all the benchmarks it tested, with exceptional performance in software engineering, knowledge work, vision and scientific research.

Safeguards and Mythos 5

Releasing a model this capable carries risk: without safeguards, Anthropic says, Fable 5's abilities in areas such as cybersecurity could be misused to cause serious damage. Queries on some topics will instead be answered by the next-most-capable model, Claude Opus 4.8. The safeguards were tuned conservatively so the release could be both safe and quick; they will sometimes catch harmless requests, but trigger in under 5% of sessions on average.

Alongside it, the company is releasing Claude Mythos 5 for a small group of cyberdefenders and infrastructure providers. It is the same underlying model with safeguards lifted in some areas, deployed initially through Project Glasswing with the US government as an upgrade to Claude Mythos Preview, and Anthropic calls it the strongest cybersecurity model in the world. A separate trusted access programme for biology will give researchers Fable 5 with biology and chemistry safeguards removed but cyber safeguards intact.

Benchmarks and research

Anthropic reports unusually strong results on long-running work. Stripe said Fable 5 compressed months of engineering into days, performing a codebase-wide migration across a 50-million-line Ruby codebase in one day — a task it says would have taken a team more than two months by hand. On Cognition's FrontierCode evaluation, which tests whether a model meets production-codebase standards on difficult coding tasks, Fable 5 scores highest among frontier models even at medium effort. Hebbia's Finance Benchmark for senior-level reasoning also placed it first, and IMC said it passed the firm's trading-analysis evaluations nearly across the board.

In vision, the model can pull precise numbers out of detailed scientific figures and rebuild a web app's source code from screenshots alone, and it beat Pokémon FireRed with a minimal, vision-only harness where earlier Claude models struggled even with extra tools. In internal research, Mythos 5 accelerated parts of drug design by around ten times and matched or beat skilled human operators with protein design and bioinformatics tools but no human assistance; nine of 14 protein targets produced strong drug-design candidates. In genomics it worked largely autonomously for over a week and outperformed a model recently published in Science despite being 100 times smaller.

Price and availability

Both models cost $10 per million input tokens and $50 per million output tokens, less than half the price of Claude Mythos Preview. Fable 5 is available everywhere; Mythos 5 stays limited to Glasswing partners and, soon, selected biology researchers. On subscription plans, Fable 5 is included at no extra cost through June 22 on Pro, Max, Team and seat-based Enterprise plans, after which it will require usage credits.