Anthropic released Claude Fable 5 today — the first publicly available version of its "Mythos-class" model and, the company says, the most powerful model it has ever made generally available. This isn't a typical Opus refresh. Fable 5 is Mythos with safeguards added — the model Anthropic was previously gating to enterprise customers and partners through Project Glasswing. Here's the SMB-relevant version of what changed.

Measurably better — and built for longer-running work

Fable 5 scores more than 10% higher than Claude Opus 4.8 on several benchmarks. The more interesting claim is the autonomy one. Anthropic says Fable 5 can "work for days at a time" inside an agent runtime like Claude Code or Claude Managed Agents — planning across stages, delegating to sub-agents, and checking its own work. If your SMB use case is long-running research, multi-step automation, or anything chained across tools, that step up matters more than the headline benchmark number.

Free in your Claude paid plan — but only until June 22

Through June 22, Fable 5 is included at no extra cost in Pro, Max, Team, and seat-based Enterprise Claude plans. On June 23, Anthropic will pull it from those plans and shift to usage-based credits. Translation: you have about two weeks to test it on your real workflows for free before you have to decide whether the upgrade is worth the cost.

The pricing is double Opus 4.8

Once the free window closes, Fable 5 costs $10 per million input tokens and $50 per million output tokens — exactly twice what Opus 4.8 costs. For most SMB chat-style use, that's still small money on an absolute basis. For high-volume agent workflows that consume tokens at scale — overnight automations, large document reviews, recurring agent chains — that doubled cost is a real budget conversation, not a rounding error.

The safety fallback is unusual and worth understanding

Fable 5 ships with safeguards that block responses in high-risk areas like cybersecurity, biology, and chemistry. When the safeguards trigger — Anthropic says less than 5% of sessions on average — the response comes from Claude Opus 4.8 instead. For most SMB use this is invisible. If your business operates in one of those domains, expect Fable 5 to quietly defer to Opus 4.8 some of the time, and plan around that.

Your next step

Try Fable 5 this week while it's free. Pick the one or two workflows where Opus 4.8 has been just-not-quite-there — long research threads, complex agent chains, multi-step document analysis. Run them through Fable 5 and compare. By June 22 you'll either know it's worth the doubled cost for those specific jobs or you'll keep using Opus 4.8 without regret. Don't decide on the price tag in the abstract; decide on the workflow.