Anthropic releases Claude Fable 5, its first Mythos-class model
Fable 5 tops agentic coding leaderboards and introduces a new safety tier system that separates general availability from approved-organization access. Demo content.
Anthropic confirmed Claude Code gets Fable 5 as the default model for Max subscribers starting this week.
Anthropic has launched Claude Fable 5, the first model in its new Claude 5 family and the debut of a Mythos-class tier positioned above Opus in capability.
Two names, one model
Fable 5 and Mythos 5 share the same underlying model. The difference is access policy: Fable 5 is generally available with additional safety measures around dual-use capabilities, while Mythos 5 is offered without those measures to approved organizations only.
Benchmarks
- ▸State-of-the-art results on agentic coding evaluations
- ▸Long-horizon task coherence measured in hours, not minutes
- ▸Improved refusal calibration on ambiguous security requests
"The tier split acknowledges something the industry has danced around: the same capability profile is safe in one deployment context and not in another." — independent safety researcher
The launch intensifies the frontier race days after OpenAI's GPT-5.2 release, with both labs now shipping models that dynamically manage their own reasoning depth.
Source: Anthropic
