Fable 5.1 is generally available through Anthropic's API and cloud platforms, while Mythos 5.1 uses the same underlying model with fewer restrictions in sensitive areas such as cybersecurity and biology. Mythos access is limited to vetted organizations.
Stronger coding and agent performance
Fable 5.1 shows a significant improvement in terminal-based agent tasks:
- Terminal-Bench 4.0: 55.8%, up from 42.0% for Fable 5;
- Mythos 5.1: 60.9%;
- Terminal-Bench-Science 0.1: Fable 5.1 scores 52.6%, versus 24.7% for Fable 5.
The models are designed for long-running workflows where an agent can inspect repositories, execute terminal commands, modify files, run tests, analyze errors and continue working without constant developer input.
Anthropic says Fable 5.1 is also better at root-cause debugging instead of applying superficial fixes.
1M-token context and 128K output
Fable 5.1 supports:
- 1 million tokens of context;
- 128,000 tokens maximum output;
- text and image input;
- adaptive reasoning;
- reliable knowledge cutoff of June 2026.
The large context window is particularly useful for coding agents working with entire repositories, documentation, schemas and API specifications.
API pricing
Fable 5.1 costs:
| Usage | Price per 1M tokens |
| Input | $10O |
| Utput | $50 |
| Cache read | $0.25 |
Input and output prices remain unchanged from Fable 5, but cache reads are 75% cheaper. Anthropic estimates typical workloads can cost around 25% less, while highly agentic applications that repeatedly reuse cached context could save approximately 45%.
This is important for coding agents because repository code and documentation can be cached instead of being processed as new input on every model call.
API availability
Developers can access Fable 5.1 through the Claude API using: claude-fable-5-1
The model is also available through Amazon Web Services, Google Cloud and Microsoft Foundry.
Fable 5.1 vs Mythos 5.1
The models share the same underlying architecture but differ in safeguards. Fable 5.1 is intended for general development and commercial applications. Mythos 5.1 provides broader capabilities for cybersecurity and life-sciences research but is restricted to approved organizations.
Anthropic says its updated cybersecurity safeguards also produce 60% fewer false positives, reducing unnecessary blocks for legitimate development and defensive-security tasks.
What changes for developers
Fable 5.1 is primarily an upgrade for developers building autonomous agents rather than simple chat applications. The key improvements are stronger coding and terminal performance, 1M-token context, long-running agent execution and cheaper context caching.
The practical result is that developers can build agents that keep large codebases in context, repeatedly use the same cached project data and perform multi-step development tasks with less human intervention and lower total API cost.
