The short version
Claude Fable 5.1 is the direct successor to Claude Fable 5, positioned by Anthropic as "the world's most advanced models for coding and knowledge work"[1]. The unusual part is that Fable 5.1 and Mythos 5.1 are the same model: Fable 5.1 is the generally available version, while Mythos 5.1 is gated behind trusted-access programs and tuned for cybersecurity and the life sciences[1]. For most readers, the practical question is whether the step up from Fable 5 (and from the cheaper Opus 5 tier) is worth switching for.
Price: cheaper than Fable 5
Anthropic says Fable 5.1 will cost an estimated 25% less than Fable 5 for typical workloads wherever usage is billed by token, driven by lower pricing on cache reads (inputs already processed and stored). For highly agentic work, the savings can reach roughly 45%[1]. That's the reverse of the usual pattern — more capable, yet cheaper — and it's one reason early partners like Cognition moved Opus 5 traffic to Fable 5.1 on launch day[1].
Benchmarks vs the field
| Benchmark | Fable 5.1 | Fable 5 | Opus 5 | GPT-5.6 Sol |
|---|---|---|---|---|
| Terminal-Bench-Science 0.1 (scientific research) | 52.6% | 24.7% | 29.0% | 22.4% |
| Terminal-Bench 4.0 (agentic coding) | 55.8% | 42.0% | 52.3% | 37.3% |
| GDPval-AA v2 (knowledge work) | 1853 | 1723 | 1824 | 1711 |
| OSWorld 2.0 (computer use, partial) | 77.9% | 72.9% | 75.4% | — |
| Humanity's Last Exam (with tools) | 65.0% | 63.8% | 63.6% | — |
| AutomationBench (business workflows) | 31.4% | 17.1% | 26.9% | 19.6% |
| CursorBench 3.2.0 (agentic coding) | 73.4% | 70.5% | 70.0% | 67.2% |
Selected results from Anthropic's Claude Fable 5.1 model announcement, September 2026[1].
The largest jump is in long-horizon scientific research — 52.6% versus 24.7% for Fable 5 on Terminal-Bench-Science 0.1 — which is why Anthropic frames these models partly as a preview of AI's role in science[1]. In day-to-day coding, the gains are more incremental but consistent: ahead of Opus 5 on Terminal-Bench 4.0 and CursorBench, and roughly doubling Fable 5 on AutomationBench[1].
Effort levels and defaults
Fable 5.1 keeps the configurable effort system from the 5.x line, but defaults higher: High effort in Claude Code, and Medium in Claude Cowork and Claude.ai[1]. At Low or Medium effort, Anthropic says Fable 5.1 already matches or beats Fable 5's results at much lower cost[1] — useful if you're trying to keep spend flat while upgrading quality.
Privacy and data retention
Anthropic is pairing the launch with a new Enterprise Frontier Safeguards (EFS) system, which it says delivers "complete privacy" — the same as a zero data retention policy — while still blocking adversarial use. Under EFS, data is stored in cloud infrastructure controlled entirely by the customer, not Anthropic[1]. EFS rolls out to enterprise customers in phases beginning later this fall; until then, eligible customers can use Fable 5.1 with zero data retention[1].
Safeguards: fewer false positives
Anthropic says its newest safeguards block 60% fewer false positives in cybersecurity, partly because Fable 5.1 can now be used to discover software vulnerabilities — though not to develop exploits for them[1]. In biology, a new access program developed with the US government governs Mythos 5.1's advanced capabilities, with enrollment for scientists expected soon[1].
What it means for small business & creators
- Lower bills for heavy users. The cache-read price cut matters most if you run long agents or re-feed the same documents repeatedly[1].
- No new plan needed. Fable 5.1 is a model upgrade inside Claude — it slots into the Claude.ai model picker and Claude Code for existing subscribers[1].
- Better long-running work. Early partners report fewer "loses the plot" failures on multi-hour, unattended tasks — relevant if you use Claude for research or coding agents[1].
The caveats
Mythos 5.1 is not something most people can access — it's restricted to trusted-access programs, so the "Mythos" name shouldn't be read as a consumer product[1]. The stated price cut is an estimate tied to token-billed usage and cache reads, so your actual savings depend on workload mix[1]. And as always with a brand-new flagship, expect the model picker and defaults to shift as Anthropic tunes rollout — check which model your account is actually using.
Bottom line
Verdict: If you're already on Fable 5 or paying for Opus 5 as a daily driver, switch — Fable 5.1 is the rare update that's both more capable and cheaper. If you're on a lower Claude tier and your work is light, there's no rush; the upgrade matters most for coding, research, and long-running agentic tasks.
Frequently asked questions
Is Claude Fable 5.1 more expensive than Fable 5?
No — Anthropic estimates it will be about 25% cheaper for typical token-billed workloads, and up to ~45% cheaper for highly agentic work[1].
What's the difference between Fable 5.1 and Mythos 5.1?
They're the same model with different safeguards. Fable 5.1 is generally available; Mythos 5.1 is gated behind trusted-access programs for cybersecurity and life-sciences work[1].
Where can I use Fable 5.1?
It's available through Claude (Claude.ai), Claude Code, and Claude Cowork, defaulting to High effort in Claude Code and Medium elsewhere[1].
Sources
- Anthropic — "Claude Fable 5.1 and Mythos 5.1" — anthropic.com
- Anthropic — Newsroom — anthropic.com
- Claude Help Center — Release notes (September 1, 2026) — support.claude.com