Claude Fable 5
A versatile model, balanced across most use cases.
Overview
Claude Fable 5, released June 9 2026, is Anthropic's first publicly available "Mythos-class" model — a tier above the Opus line. Priced at $10/$50 per million tokens (double Opus 4.8), 1M context with no long-context surcharge, 128K max output. It reaches 80.3% on SWE-Bench Pro versus 69.2% for Opus 4.8. Adaptive thinking is permanently on and raw chain-of-thought is never returned. Notable caveats: mandatory 30-day data retention (no zero-retention option), and safety classifiers reroute under 5% of sessions to Opus 4.8 on cyber, biology and chemistry topics. Fable 5 and Mythos 5 share the same underlying model, Fable being the safety-hardened general release.
Skill profile
Not disclosed
Strengths
- World-class on Arena
- Very long context
Limitations
- High output cost
- No native GDPR guarantee
Who is it for
- your budget is tight
- you handle sensitive EU data
Access & availability
Key specifications
Estimate your monthly cost
Per-token API pricingFor the same usage
- Muse Spark 1.1$68-90%
- Claude Opus 5$327-50%
- GPT-5.6 Sol$363-45%
Indicative estimate based on standard API rates (excluding caching, batch and volume discounts). Always check the official pricing before committing.
Privacy
Advanced data · for expertsArchitecture, modalities, detailed cost, full benchmarks▾
| Arena Elo | 1509 |
| MMLU | — |
| GPQA | — |
| HumanEval | — |
| SWE-Bench | — |
| MATH | — |