On 9 June 2026, Anthropic announced Claude Fable 5 and Claude Mythos 5. Anthropic exposed the same underlying capability as safeguarded Fable for general use and less-restricted Mythos for trusted high-risk work.
This release brief was checked against first-party material on 10 August 2026. The date above is the public announcement date, not the date a repository was created or a third-party provider added the model. Where access or weights arrived later, that distinction is recorded below.
Release record
| Field | Verified detail |
|---|---|
| Announcement | 9 June 2026 |
| Availability or weight release | 9 June; suspended 12 June and redeployed 1 July 2026 |
| Release type | frontier model with general and trusted-access variants |
| Access | Fable in Claude products and API with fallbacks; Mythos for approved defenders and research partners |
| Architecture | one undisclosed Mythos-class model served through two safeguard and access configurations |
| Maximum stated context | millions of tokens across long-running tasks through context and persistent memory mechanisms |
What changed
The release made safeguards observable in ordinary task routing. Anthropic said Fable’s classifiers triggered in fewer than 5% of sessions on average and routed some requests to Opus 4.8. Access was then suspended on 12 June and restored on 1 July after an export-control issue, demonstrating that availability is part of model risk. Benchmarking Fable without recording fallbacks can therefore misattribute which model completed the task.
The practical comparison is therefore not simply whether Claude Fable 5 and Claude Mythos 5 has the largest headline score. Teams need to ask whether its architecture, access terms, latency, tool behaviour and evaluation setup match the workload they actually intend to run. A model can lead one harness while losing on cost, refusal behaviour, multilingual quality or repeatability in another.
Benchmarks worth retaining
| Evaluation | Reported result | How to read it |
|---|---|---|
| Safety-classifier intervention | under 5% of sessions on average | General-use routing rate reported at launch |
| Drug-design acceleration | around 10× in Anthropic’s internal work | Mythos 5 tool-assisted research claim |
| Persistent-memory game result | final act reached about 3× as often as Opus 4.8 | Long-horizon memory experiment |
These are release-time results, not independently reproduced guarantees. Fable and Mythos share underlying weights but differ in restrictions and eligible users, while some life-science and cyber results come from internal tool environments. Scores should remain attached to the disclosed effort setting, agent harness, tool access, timeout, context-management policy and judge model. Moving a number into a procurement sheet without those conditions creates false comparability.
Architecture and access
Claude Fable 5 and Claude Mythos 5 is described as one undisclosed Mythos-class model served through two safeguard and access configurations with millions of tokens across long-running tasks through context and persistent memory mechanisms of stated context. Its access position at verification time is Fable in Claude products and API with fallbacks; Mythos for approved defenders and research partners. That wording matters: open weights, source-available weights, an API, a product preview and a research demonstration give adopters very different rights and different levels of reproducibility.
Before deployment, record the exact model identifier or checkpoint, inference stack, quantisation, reasoning setting, region, price schedule and supplier terms. If the release uses a custom licence, read the licence itself rather than relying on the word “open” in launch copy. If it is API-only, preserve the dated documentation and change-notice route because the served snapshot can change without a downloadable artefact.
What an evaluation should test next
For Claude Fable 5 and Claude Mythos 5, a credible internal gate should include:
- a frozen set of representative tasks with pass, fail and abstain criteria;
- a matched baseline using the same tools, timeout, prompt budget and reviewer rubric;
- repeated runs to expose variance rather than reporting a single best attempt;
- latency, token use and total task cost alongside task success;
- adversarial, multilingual and long-context cases relevant to the real deployment; and
- rollback evidence showing the previous model can be restored safely.
The wider model change-control guide explains how to keep model, prompt, tool and corpus changes reconstructable. The AI dependency inventory guide covers the release and supplier records needed after deployment.
AIEngine verdict
This was a landmark capability-and-access release. Any evaluation must log model routing, refusal and fallback rates, and it must treat the June suspension and July redeployment as part of the operational history.
This is a launch assessment, not a certification. Benchmark leadership is useful evidence of where to test; it is not authorization to place the model in a high-impact workflow without domain evaluation, security review and an accountable owner.
Primary sources
- Anthropic Fable 5 and Mythos 5 announcement
- Anthropic redeployment notice
- Claude Fable and Mythos system card
Image provenance
Hero image: Anthropic official release artwork. The locally served WebP is a crop of the first-party release or model-card asset recorded in the repository provenance manifest.
