Claude Fable 5
Anthropic Claude Fable 5 for long-running coding, research, and knowledge-work tasks. Anthropic retains prompts and outputs for 30 days under its Covered Models policy; Zero Data Retention is not available.
- Reasoning
- Vision
- Native PDF input
- Tool Calling
- Structured Output
Added Jun 9, 2026
Pricing
Auto routing · per 1M tokens- Input
- $10.00
- Output
- $50.00
- Cache read
- $1.00
Specifications
- Context window
- 1M
- Max output
- 128K
- Avg output (7d)
- 2K tokens
- Longer than 92% of models
Benchmarks
Benchmarks
Sourced from Artificial Analysis.
Intelligence Index
49.6
Coding Index
76.5
Agentic Index
50.7
Agentic work
AutomationBench-AA
Workflow automation with guardrail penalties
54.1%
Better than 73% of models compared
Harvey LAB-AA
Legal agentic work criterion pass rate
93.6%
Better than 93% of models compared
AA-Briefcase
Agentic knowledge work (Elo)
1539 Elo
Better than 90% of models compared
GDPval-AA v2
Economically valuable tasks (Elo)
1613 Elo
Better than 91% of models compared
Document reasoning
GDP.pdf
Professional PDF reasoning: all-pass rate
24.0%
Better than 82% of models compared
AA-LCR v1.1
Long context reasoning with updated grading
82.3%
Better than 93% of models compared
MLCR-AA
Medical long-context reasoning
64.4%
Better than 96% of models compared
Reasoning
HLE
Humanity's Last Exam
55.5%
Better than 99% of models compared
IFBench
Instruction-following benchmark
63.5%
Better than 74% of models compared
CritPt
Research-level physics reasoning
28.6%
Coding
Terminal-Bench v4.0
Practical coding and terminal tasks
42.4%
Better than 88% of models compared
SciCode
Python programming for scientific computing
61.0%
Better than 98% of models compared
Knowledge
AA-Omniscience Accuracy
Proportion of correctly answered questions
65.3%
AA-Omniscience Hallucination Rate
Rate of incorrect answers among non-correct responses
63.6%
Legacy benchmarks
GPQA Diamond (legacy)
Graduate-level scientific reasoning
92.6%
Better than 94% of models compared
Terminal-Bench Hard (legacy)
Agentic coding and terminal use
62.9%
Better than 99% of models compared
T²-Bench Telecom (legacy)
Conversational AI agents in dual-control scenarios
98.5%
Better than 99% of models compared
AA-LCR (unversioned / legacy)
Long context reasoning evaluation
82.3%
Better than 93% of models compared
GDPval-AA (unversioned / legacy)
Economically valuable tasks
55.6%
Last updated Oct 10, 2026
Artificial AnalysisProviders
Auto routing is available for this model. Explicit provider selection is not available.
Loading provider options…