Anthropic on Monday set the price for its newly launched Claude Fable 5 and Claude Mythos 5 models at $10 per million input tokens and $50 per million output tokens, which it said is less than half the rate it had previously charged for the earlier Claude Mythos Preview.
The pricing applies to both versions of the model. Fable 5, the safer version released to general users, and Mythos 5, the restricted version with cybersecurity safeguards lifted for vetted partners, share the same underlying model and the same per-token rate. They are accessible via the Claude API and Anthropic’s consumer and enterprise interfaces. The new pricing reflects a continued downward trend in frontier AI access costs.
Mythos Preview, released in April under Anthropic’s Project Glasswing programme, had been the company’s most expensive model. By bringing Mythos-class capabilities to a lower price point on a wider release, Anthropic appears to be following the broader industry pattern, where successive generations of frontier models have tended to undercut their predecessors even as their capabilities have grown.
For subscription users on Anthropic’s existing plans — Pro, Max, Team and seat-based Enterprise — Fable 5 will be included at no additional cost until June 22, 2026. From June 23, continued use will require purchased usage credits. Anthropic said the staggered rollout is intended to manage what it expects to be high and unpredictable demand. The company added that it intends to restore Fable 5 as a standard part of subscription plans once capacity allows, though it has not specified a timeline. Mythos 5 access remains restricted.
Existing Project Glasswing partners, including the US government and a small group of cybersecurity organisations and critical software infrastructure providers, will be able to upgrade from Mythos Preview to Mythos 5. Anthropic said it plans to expand access through a wider trusted access programme, including a separate track for biomedical research, where biology and chemistry safeguards will be lifted for vetted scientists. For developers, both models are available on the Claude API from today.
The pricing comes amid intensifying competition in the frontier AI market, with rivals including OpenAI and Google DeepMind continuing to roll out new models at similar capability tiers. Anthropic’s announcement positioned Fable 5 as “state-of-the-art on nearly all tested benchmarks of AI capability”, though independent verification of those claims by users and benchmark organisations is likely to follow over the coming weeks.
-
Sleeping Problem: Sleep is interrupted again and again at night? These 5 habits may be responsible
-
Amazing Benefits of Traveling Alone: Experience Before 30
-
What is bemotrizinol? Korean sunscreen ingredient gets FDA approval
-
A woman sitting outside her house in Dabra was crushed by a speeding car, the woman died in the accident, the driver fled from the spot.
-
Chinese celebrity dog with 1.5 million followers stolen, sold for $25 and eaten
