Anthropic
Fable 5
- Context
- 1M
- Output tokens
- 128k
- Input / output per 1M tokens
- $10 / $50
Anthropic releases Claude Fable 5 and Claude Mythos 5 (새 탭에서 열림)
Fable 5 returns stop_reason refusal as an ordinary response, so integrations need a path for it.
Fable 5 ships, vanishes under the first export control on model weights, and returns. Plus Sonnet 5, Nemotron 3, and the stateless MCP SDKs.
Anthropic
Fable 5
Fable 5 returns stop_reason refusal as an ordinary response, so integrations need a path for it.
The workhorse price point for agent fleets, with cyber safeguards on by default.
Regarded at release as the strongest open-weights American model and a serious option for self-hosted agent backends.
Breaking changes included: Python's FastMCP becomes MCPServer and the TypeScript SDK splits into server and client packages.
Codex is becoming a general-purpose agent harness, not just a coding tool.
Cross-vendor agent interop, landing inside a mainstream IDE.
acceptEdits now always prompts before writing executable-adjacent files like .zshenv or .npmrc.
If you depend on a single frontier model, this month taught you to build fallback paths.
Worth reading for how refusal behavior at the frontier is likely to be standardized.
June API roadmaps had to plan around staged availability.
No mandatory licensing, but it shapes when new frontier models reach the API you build on.
Directly applicable if you manage long-horizon agent context.
Useful calibration for any claim of autonomous research.