Anthropic launched two fashions on June 9, 2026: Claude Fable 5 and Claude Mythos 5. Each belong to a tier known as “Mythos-class.” This tier sits above the Opus class in functionality. Fable 5 is the model claimed to be made protected for common use. Mythos 5 is identical mannequin with some safeguards lifted, saved in restricted launch.
Claude Fable 5 and Mythos 5
Mythos-class fashions are a tier of Claude fashions. They sit above the Opus class in functionality. The primary was Claude Mythos Preview, launched in April by means of Mission Glasswing.
Fable 5 and Mythos 5 share the identical underlying mannequin. The distinction is the safeguards. Fable 5 ships with security classifiers for common use. Mythos 5 has some classifiers eliminated and stays in restricted launch.
The names mirror this cut up. “Fable” comes from the Latin fabula, “that which is instructed.” That is akin to the Greek mythos. The safeguards distinguish the 2 fashions, so that they carry completely different names.
Anthropic workforce calls Fable 5 its most succesful extensively launched mannequin. It targets demanding reasoning and long-horizon agentic work. Anthropic states Fable 5’s capabilities exceed any mannequin it has made typically obtainable.
Each fashions help a 1M token context window by default. They permit as much as 128k output tokens per request. Pricing is $10 per million enter tokens and $50 per million output tokens. That’s lower than half the value of Claude Mythos Preview.
The Functionality Case
Anthropic experiences Fable 5 is state-of-the-art on almost all examined functionality benchmarks. It reveals robust outcomes throughout software program engineering, information work, imaginative and prescient, and scientific analysis. The longer and extra complicated the duty, the bigger its lead over Anthropic’s different fashions.
On software program engineering, Stripe examined Fable 5 throughout early entry. The mannequin carried out a codebase-wide migration in a 50-million-line Ruby codebase. In accordance with Stripe: this took sooner or later. By hand, a workforce would have wanted over two months.
Fable 5 can also be extra token-efficient than previous Claude fashions. On Cognition’s FrontierCode analysis, Fable 5 scores highest amongst frontier fashions. This holds even at medium effort. The eval exams troublesome coding duties beneath production-codebase requirements.
On information work, Anthropic cites Hebbia’s Finance Benchmark for senior-level reasoning. Fable 5 posts the best rating of any mannequin there. Good points are available document-based reasoning, chart and desk interpretation, and drawback fixing.
On imaginative and prescient, Anthropic calls Fable 5 the brand new state-of-the-art. It may well extract exact numbers from detailed scientific figures. It may well rebuild an online app’s supply code from screenshots alone. It additionally wants much less scaffolding than prior fashions. Fable 5 beat Pokémon FireRed with a minimal, vision-only harness.
On reminiscence and long-context, Fable 5 stays targeted throughout tens of millions of tokens. It improves its outputs utilizing its personal notes. Within the sport Slay the Spire, persistent file-based reminiscence helped it thrice greater than Opus 4.8.
Mythos 5 carries the science claims. Inner protein design specialists accelerated elements of drug design by round ten instances. Anthropic additionally says Mythos 5 is its first mannequin to persistently produce novel scientific hypotheses. Scientists most popular its molecular biology hypotheses round 80% of the time in blinded comparisons.
Mythos 5 additionally ran novel genomics analysis over every week of largely autonomous work. It skilled a customized mannequin on single-cell knowledge spanning 138 animal species. Anthropic says that mannequin outperformed a current mannequin printed in Science, regardless of being 100 instances smaller.
How the Safeguards Work
Releasing a mannequin this succesful carries danger. With out safeguards, Fable 5’s cybersecurity capabilities could possibly be misused to trigger critical injury. Anthropic due to this fact launched Fable 5 with a brand new set of classifiers.
Classifiers are separate AI techniques. They detect potential misuse, together with jailbreak makes an attempt. They stop the principle mannequin from responding to flagged requests.
When Fable 5’s classifiers flag a request, the response is dealt with by Claude Opus 4.8 as an alternative. The lined areas are cybersecurity, biology and chemistry, and distillation. Customers are knowledgeable each time a fallback happens.
For biology and chemistry, Fable 5 falls again to Opus 4.8 on most requests for now. Anthropic cites concern that the identical dual-use queries may give uplift to malicious actors. It plans a trusted entry program for biology, giving accredited researchers Fable 5 with out these safeguards.
Anthropic tuned these safeguards conservatively. They are going to generally catch innocent requests. On common, they set off in lower than 5% of classes. Anthropic says greater than 95% of Fable classes contain no fallback in any respect. For these classes, Fable 5’s efficiency successfully matches Mythos 5.
Anthropic red-teamed the classifiers extensively. An exterior bug bounty produced no common jailbreaks in over 1,000 hours. A common jailbreak lets a person work together with the mannequin as if its safeguards have been absent. Anthropic notes the UK AISI made progress towards one in a quick testing window.
Mythos 5 is identical mannequin with cyber safeguards lifted. Anthropic describes it as having the strongest cybersecurity capabilities of any present mannequin. It’s deployed by means of Mission Glasswing in collaboration with the US authorities.
Use Instances
These capabilities map to a number of concrete workflows for technical groups:
- Giant-scale code migration: Lengthy-horizon coding fits massive refactors and cross-repo migrations. The Stripe instance reveals this at a 50-million-line scale.
- Agentic coding pipelines: Fewer turns and token effectivity assist multi-step agent runs. GitHub reported autonomy and reliability on complicated, long-horizon coding duties.
- Finance and analytics work: Robust doc and chart reasoning fits senior-level monetary evaluation. Hebbia and IMC cited good points on reasoning and trading-analysis duties.
- Imaginative and prescient-to-code duties: Rebuilding supply from screenshots fits front-end reconstruction and determine extraction. The vision-only harness reduces tooling overhead.
- Lengthy-running analysis brokers: Persistent reminiscence throughout tens of millions of tokens fits multi-day analysis loops. Mythos 5 ran novel genomics work over every week of largely autonomous work.
Comparability Desk: Fable 5 vs. Mythos 5 vs. Opus 4.8
| Attribute | Claude Fable 5 | Claude Mythos 5 | Claude Opus 4.8 |
|---|---|---|---|
| Mannequin tier | Mythos-class | Mythos-class | Opus class |
| Underlying mannequin | Identical as Mythos 5 | Identical as Fable 5 | Opus 4.8 |
| Availability | Usually obtainable | Restricted (Mission Glasswing) | Usually obtainable |
| Security classifiers | Energetic (cyber, bio/chem, distillation) | Cyber safeguards lifted | Opus-level safeguards |
| Fallback goal | Falls again to Opus 4.8 | Not relevant | Not relevant |
| API mannequin ID | claude-fable-5 |
claude-mythos-5 |
(current Opus ID) |
| Context window | 1M tokens default | 1M tokens default | Per Opus specs |
| Max output | 128k tokens/request | 128k tokens/request | Per Opus specs |
| Enter worth (per 1M) | $10 | $10 | (per Opus pricing) |
| Output worth (per 1M) | $50 | $50 | (per Opus pricing) |
| Pondering mode | Adaptive solely, at all times on | Adaptive solely, at all times on | Configurable |
| Knowledge retention | 30-day (Lined Mannequin) | 30-day (Lined Mannequin) | Normal choices |
Be aware: Particular Opus 4.8 specs and pricing should not detailed within the Fable 5 launch sources. The desk marks these cells accordingly.
Key Takeaways
- Fable 5 and Mythos 5 share one underlying mannequin; safeguards are the one distinction.
- Anthropic experiences Fable 5 is state-of-the-art on almost all examined functionality benchmarks.
- Fable 5 classifiers fall again to Opus 4.8 and set off in beneath 5% of classes.
- Each fashions supply a 1M token context window at $10 enter and $50 output per million tokens.
- Mythos 5 stays restricted to Mission Glasswing; Fable 5 is mostly obtainable throughout main platforms.
| Src | Writer / Publish | Theme | Attain | Sentiment |
|---|
Try the Technical particulars, Docs and Sentiments Analytics from the AI Group. Additionally, be happy to observe us on Twitter and don’t overlook to affix our 150k+ ML SubReddit and Subscribe to our E-newsletter. Wait! are you on telegram? now you possibly can be a part of us on telegram as nicely.
Have to associate with us for selling your GitHub Repo OR Hugging Face Web page OR Product Launch OR Webinar and so on.? Join with us
