Claude Fable 5: Skills Setup, Pricing, and What Happened
Claude Fable 5 launched June 9, restricted June 12. What happened, SKILL.md setup, pricing, and availability.
Quick Answer: Claude Fable 5 launched June 9, 2026 as the strongest coding model available. On June 12, Anthropic restricted access globally following U.S. export controls. Where available, it leads SWE-Bench at 88.6% and is accessible through Claude Code and the API. SKILL.md skills work with Fable 5 the same as any Claude model.
Fable 5 is the most powerful coding model Anthropic has released. It is also the most complicated to access. Here is what happened, where it stands, and how to use it with SKILL.md skills.
What Fable 5 is
Fable 5 is Anthropic's June 2026 release in the Claude 5 family. It leads SWE-Bench Verified at 88.6%, SWE-Bench Pro at 80.3%, and Terminal-Bench 2.1 at 88.0%. Anthropic reports that Stripe used it to migrate a 50-million-line codebase in a single day. It is the first model to break 90% on several core analytics benchmarks.
Pricing: $10 per million input tokens, $50 per million output tokens. Double the cost of Opus 4.8 but significantly more capable for complex coding tasks.
Recommended skills
Price Competitor Scan
by Edric Vale
Competitive price intelligence: track rival pricing, detect changes, and recomme…
Agent Evaluation & Regression Suite — Score Skills Before You Ship
by Vivek K
Designs a scored eval suite for an AI agent or SKILL.md skill, measures trigger…
Hooks & Settings for Claude Code
by Markus Isaksson · 23
Master Claude Code's settings hierarchy and hook framework to automate workflows…
What happened on June 12
Three days after launch, Anthropic restricted Fable 5 and Mythos 5 globally following a U.S. government export control directive. The specifics vary by region and plan, but the result is that Fable 5 is not as freely available as Anthropic initially intended.
This makes Fable 5 the second frontier model in June 2026 to have its availability shaped by government action (GPT-5.6 being the other). The pattern suggests that future frontier model launches will increasingly involve regulatory coordination.
Using SKILL.md skills with Fable 5
SKILL.md skills are model-agnostic. They define what the agent should do. The model determines how well it does it. If you have access to Fable 5 through Claude Code, your existing skills from the Agensi marketplace work without any changes.
In practice, Fable 5's stronger reasoning means skills produce better output. A code review skill running on Fable 5 catches more subtle bugs than the same skill on Sonnet 4.6, because the underlying model is more capable.
For Claude Code setup, see how to install skills in Claude Code.
If you do not have Fable 5 access
Claude Opus 4.7 remains an excellent coding model and is widely available through Claude Pro ($20/month). Opus 4.7 still wins 67% of blind code reviews and has the most mature SKILL.md ecosystem. For most development work, the difference between Opus 4.7 and Fable 5 is meaningful but not transformative.
Browse 2,000+ skills that work with any Claude model on the Agensi marketplace.