‹ Back

Anthropic Launches Claude Fable 5, Its Most Powerful Public AI

 

Anthropic releases Claude Fable 5 and Claude Mythos 5, its first Mythos-class models — one public, one restricted to vetted partners.

 

 

Anthropic Opens Mythos-Class AI to General Users for First Time

 

Anthropic launched Claude Fable 5 on June 9, 2026 — a Mythos-class model the company has made safe for general use. The release marks the first time the San Francisco-based AI company has made a model from its top-tier Mythos category accessible to the public, representing a notable shift from the controlled rollout that defined the class since its introduction two months earlier.

 

Fable 5's capabilities exceed those of any model Anthropic has previously made generally available. The company states it is state-of-the-art on nearly all tested benchmarks of AI capability, with exceptional performance in software engineering, knowledge work, vision, scientific research, and other areas. The longer and more complex the task, the larger Fable 5's lead over Anthropic's other models.

 

Alongside Fable 5, Anthropic simultaneously launched Claude Mythos 5 — the same underlying model but with safeguards lifted in some areas — for a small group of cyberdefenders and infrastructure providers. The two models share identical weights, capabilities, and pricing; what distinguishes them is access control, determined by the user's verified identity and affiliation rather than by any consumer-facing selection.

 

Safeguards Built Into Fable 5 to Contain High-Risk Queries

 

Releasing a model this capable carries risks. Without safeguards, Fable 5's capabilities in areas like cybersecurity could be misused to cause serious damage. Anthropic has therefore launched the model with safeguards that mean queries on some topics will instead receive a response from its next-most-capable model, Claude Opus 4.8. To release the model both safely and quickly, the company tuned these safeguards conservatively — they will sometimes catch harmless requests, though they trigger, on average, in less than 5% of sessions.

 

The safeguards operate through classifiers: separate AI systems that detect potential misuse, including jailbreak attempts, and prevent Fable 5 from responding. When the classifiers detect a request related to cybersecurity, biology and chemistry, or distillation, the response is automatically handled by Claude Opus 4.8 instead. Users are informed whenever this fallback occurs.

 

On cybersecurity, Anthropic extensively red-teamed its classifiers to test robustness against jailbreaks. An external bug bounty produced no universal jailbreaks in over 1,000 hours of testing. External red-teaming organizations also failed to find any universal jailbreaks on long-form agentic tasks — although the UK AI Security Institute made progress toward one within a brief initial testing window. Anthropic acknowledged that completely preventing universal jailbreaks is likely impossible, and stated its aim is to make any remaining jailbreaks sufficiently costly that it can detect and prevent them before they are deployed at scale.

 

On biology and chemistry, Anthropic tested Mythos 5's ability to complete a challenging step in designing adeno-associated viruses (AAVs), a component used in gene therapy delivery. Mythos-class models outperformed sophisticated protein language models using biological reasoning alone — without explicit task-specific training — on unpublished experimental candidates developed by Dyno Therapeutics. Anthropic cited this dual-use risk as the basis for broad biological safeguards, which it intends to narrow as its classifiers improve.

 

A third category of safeguards addresses distillation. Anthropic has previously identified large-scale attempts to extract Claude's capabilities to train competing models. Requests flagged as distillation attempts will fall back to Opus 4.8.

 

Performance in Software Engineering, Science, and Knowledge Work

 

During early testing, Stripe reported that Fable 5 compressed months of engineering into days. In a 50-million-line Ruby codebase, the model performed a codebase-wide migration in a day that would otherwise have taken a whole team over two months by hand. Cursor CEO and co-founder Michael Truell stated that Fable 5 is the state-of-the-art model on CursorBench and has opened up a class of long-horizon problems that were previously out of reach.

 

On knowledge work, Fable 5 achieved the highest score of any model on Hebbia's Finance Benchmark for senior-level reasoning, with substantial gains in document-based reasoning, chart and table interpretation, and problem solving. IMC noted that Fable 5 performed at or near the top across its trading-analysis evaluations, including factual lookup, conceptual reasoning, root-cause analysis, and expected-value analysis.

 

In life sciences research using Mythos 5, Anthropic's internal protein design experts reported accelerating aspects of the drug design process by around ten times. In one instance, Mythos 5 — operating with protein design and bioinformatics tools but without human assistance — matched or exceeded the performance of skilled human operators across the full set of tasks typically completed by a scientist, including choosing binding sites, selecting and running protein design tools, and recovering from failures. Nine of 14 protein targets yielded strong drug design candidates.

 

Mythos 5 also conducted novel genomics research over more than a week of largely autonomous work. It assembled single-cell data spanning 138 animal species and designed and trained a custom machine learning model to identify cells performing the same functional role across distantly related organisms. The model Mythos 5 trained outperformed a recently published model from the journal Science, while being 100 times smaller. Anthropic stated it intends to publish these results.

 

Mythos 5 Access Tied to Project Glasswing and Trusted Programs

 

Anthropic launched Project Glasswing on April 7, 2026, bringing together Amazon Web Services, Apple, Broadcom, Cisco, CrowdStrike, Google, JPMorganChase, the Linux Foundation, Microsoft, NVIDIA, and Palo Alto Networks in an effort to secure critical software infrastructure. The initiative was formed in response to Claude Mythos Preview's demonstrated ability to find thousands of high-severity vulnerabilities, including in every major operating system and web browser.

 

Beginning June 9, all users with existing access to Claude Mythos Preview — including Project Glasswing cybersecurity partners — are able to upgrade to Claude Mythos 5, with cyber safeguards lifted. Anthropic stated that Mythos 5 is comparable to, or somewhat stronger than, Mythos Preview in most cases, while costing substantially less. Anthropic also announced plans for a trusted access program for biology, which will provide access to Fable 5 with biology and chemistry safeguards removed, targeting a small number of researchers from life science organizations across fundamental and translational research.

 

Pricing, Data Retention, and Availability Schedule

 

Both Fable 5 and Mythos 5 are priced at $10 per million input tokens and $50 per million output tokens — less than half the price of Claude Mythos Preview. The models carry a one-million-token context window, 128,000 maximum output tokens, and a knowledge cutoff date of January 2026.

 

Anthropic is also implementing a new data retention policy for Mythos-class models. It will require 30-day retention for all traffic on Mythos-class models on both first- and third-party surfaces. The company stated it will not use this data to train new Claude models or for any non-safety-related purpose, and has instituted new privacy protections including logging all human access and ensuring deletion after 30 days in almost all cases.

 

On the Claude API and consumption-based Enterprise plans, Fable 5 is fully available as of launch. For subscription plans, Fable 5 is included on Pro, Max, Team, and seat-based Enterprise plans at no extra cost through June 22. On June 23, the company will remove Fable 5 from those plans; use after that date will require usage credits until sufficient capacity allows Anthropic to restore it as a standard feature.