Anthropic has launched Claude Fable 5 and Claude Mythos 5, introducing a new "Effort" parameter to dynamically balance reasoning capability, latency, and cost for autonomous agentic workflows.
Anthropic released Claude Fable 5 and Claude Mythos 5, two versions of its latest frontier intelligence model designed for complex, long-horizon agentic tasks. Claude Fable 5 is generally available and includes built-in safety classifiers that redirect risky requests to Claude Opus 4.8, while Claude Mythos 5 provides raw, unsafeguarded performance exclusively to approved partners. A key feature of the release is the new "Effort" parameter, which allows developers to configure the amount of reasoning (low, medium, high, xhigh) the model performs, directly optimizing the trade-offs between performance, latency, and token cost.
Anthropic is shifting the paradigm from static conversational AI to active planning agents, signaling a transition where compute-over-reasoning and safety-driven fallback systems are first-class developer configurations.
* The "Effort" parameter formally commercializes test-time compute, allowing developers to pay only for the depth of reasoning a specific task actually requires.
* Built-in safety routing (automatically falling back to Claude Opus 4.8) highlights the challenge of securing agentic models without outright breaking application stability.
* Separating the model into public (Fable) and partner-restricted (Mythos) releases allows Anthropic to satisfy safety regulators while keeping frontier capabilities accessible to high-security research.
DISCOVERED
1h ago
2026-06-11
PUBLISHED
2h ago
2026-06-11
RELEVANCE
AUTHOR
AlphaSignalAI