Claude Mythos 5 is Anthropic's limited-release frontier model with identical capabilities to Claude Fable 5 — but without safety classifiers that can refuse requests. It succeeds Claude Mythos Preview and is available through Project Glasswing.
Mythos 5 vs Fable 5
Same 1M context, 128K max output, adaptive thinking, and API pricing ($10/$50 per million tokens). The difference is policy: Mythos 5 never applies refusal classifiers, making it suitable for teams that need unconstrained completions in controlled enterprise environments.
Who can access Mythos 5?
Not generally available. Approved customers receive access via Project Glasswing through Anthropic, AWS, or Google Cloud account teams. If you lack access, Claude Fable 5 is the public alternative with the same core intelligence plus safety layers.
When Mythos makes sense
Regulated internal pipelines where safety is enforced upstream, research environments with human review, and API products that cannot tolerate unexpected refusals mid-workflow. Not recommended for consumer-facing apps without additional guardrails.
Benchmark context
Independent orchestration systems like Sakana Fugu benchmark against Fable 5 and Mythos Preview as frontier baselines. See our full flagship comparison.
