Claude Fable 5.1

Why teams choose it

Tradeoffs to know

• Cyber offense such as exploit generation still redirects; Fable 5.1 may discover vulnerabilities but not develop exploits.
• Vendor Terminal-Bench and science scores use Anthropic's harness; reproduce before replacing Opus 5 or Fable 5.
• Default effort is high in Claude Code and medium on Claude.ai—token use can jump if you leave defaults on.

When not to use this

• Not ideal for simple tasks where cheaper models in the same lineup are good enough.
• Avoid for latency-sensitive real-time chat when raw response speed outweighs reasoning depth.
• Confirm limits, pricing, and regional availability on the provider side before committing production workloads.

Technical specs

Inputs

text, image

Outputs

text

Capabilities

long-horizon coding, agentic research, adaptive thinking, computer use, document work, tool use, long context

License

Proprietary API

Model string

claude-fable-5-1

Benchmarks

{
  "source": "https://www.anthropic.com/claude-fable-and-mythos-5-1",
  "gdpval_aa_v2": 1853,
  "hle_no_tools": 60.9,
  "automationbench": 31.4,
  "vendor_reported": true,
  "cursorbench_3_2_0": 73.4,
  "terminal_bench_4_0": 55.8,
  "osworld_2_0_partial": 77.9,
  "terminal_bench_science_0_1": 52.6
}