ModelBeat
Models/Claude Fable 5
All models
A

Anthropic: Claude Fable 5

Watch this modelCompareAnnouncement
anthropic/claude-fable-5

Claude Fable 5 is Anthropic’s proprietary AI model, released Jun 9, 2026. It scores 164 on the Epoch Capabilities Index, costs $10/M input and $50/M output tokens, and handles a 1M-token context window. Its strongest use case is coding, where it ranks in the 90th percentile of the models Model Beat tracks.

Claude Fable 5 is a Mythos-class model from Anthropic, built for autonomous knowledge work and coding.

Context
1M
Input
$10 / 1M
Output
$50 / 1M
Released
Jun 9, 2026
Modalities
textimagefile→text
Strongest atCoding top 10%Reasoning & Knowledge top 11%Agentic & Tools top 11%

Benchmarks

Scores on standardized evaluations. Higher is better — and rank shows where Claude Fable 5 lands among all models tracked on Model Beat.

88.0
Intelligence Index
Epoch AI
88th percentile of tracked models
90.0
Coding Index
Epoch AI
90th percentile of tracked models
89.0
Agentic Index
Epoch AI
89th percentile of tracked models

Reasoning

7 evals
ARC-AGI98.5%

Abstract visual reasoning

ARC-AGI-289.2%

Harder abstract reasoning

GPQA Diamond93.9%

Graduate-level scientific reasoning

Humanity's Last Exam55.5%

Frontier of human expert knowledge

SimpleBench81.9%

Common-sense trick questions

SimpleQA Verified70.7%

Factual accuracy & hallucination

WeirdML91.9%

Novel ML problem-solving

Coding

3 evals
GSO (code optimization)78.4%

Code performance optimization

SciCode61.0%

Scientific research coding

WebDev Arena1628

Human-rated web development

Math

1 evals
AIME 2024/2025100.0%

Olympiad-qualifier math

Agentic & Tools

2 evals
APEX63.6%

Multi-step agentic tasks

τ²-bench98.5%

Tool-agent-user reliability

In the news

30

Changelog

3
  • Claude Fable 5: reported GPQA Diamond score rose from 85.9% to 93.9% (+9%)benchmark · Nov 13, 2026
  • Claude Fable 5: reported APEX score rose from 45.0% to 63.6% (+41%)benchmark · Sep 22, 2026
  • Claude Fable 5: GPQA Diamond improved from 78.8% to 85.9% (+9%)benchmark · Aug 7, 2026

Compare Claude Fable 5 with

Claude Fable 5 vs GPT-6 AstraClaude Fable 5 vs Claude Opus 5.5Claude Fable 5 vs Gemini 3.8 FlashClaude Fable 5 vs Kimi K3Claude Fable 5 vs Grok 4.6

Frequently asked questions

What is Claude Fable 5?

Claude Fable 5 is a Mythos-class model from Anthropic, built for autonomous knowledge work and coding.

Who created Claude Fable 5?

Claude Fable 5 is an AI model developed by Anthropic, part of the Claude family, released Jun 9, 2026.

How much does Claude Fable 5 cost?

Claude Fable 5's list price from Anthropic is $10 per million input tokens and $50 per million output tokens.

What is Claude Fable 5's context window?

Claude Fable 5 supports a context window of up to 1M tokens.

How does Claude Fable 5 perform on benchmarks?

On standardized evaluations tracked by Epoch AI, Claude Fable 5 scores 55.5% on Humanity's Last Exam, 61.0% on SciCode, 98.5% on τ²-bench.

Is Claude Fable 5 open source?

Claude Fable 5 is a proprietary model (API access), available through its provider's API rather than as a downloadable open-weight model.

Benchmark data from Epoch AI (CC BY) and Artificial Analysis; pricing, specs & descriptions from OpenRouter. Data updated Nov 13, 2026.