Anthropic has unveiled Claude Fable 5.1 and Claude Mythos 5.1, introducing a new generation of AI models designed for coding, knowledge work and complex research tasks.
The San Francisco-based AI company says Fable 5.1 delivers significant performance improvements over its predecessor while also reducing costs for many workloads. Mythos 5.1 uses the same underlying model but comes with different safeguards and is being offered through restricted trusted-access programs.
Claude Fable 5.1 Targets More Complex Work
Anthropic is positioning Fable 5.1 for tasks that require AI systems to work through complicated problems over longer periods of time.
The company reports that Fable 5.1 scored 55.8% on Terminal-Bench 4.0, compared with 42.0% for Fable 5. On Terminal-Bench-Science 0.1, Fable 5.1 scored 52.6%, more than double the 24.7% reported for its predecessor.
Anthropic also reported improvements across knowledge work, computer use, business workflows and agentic coding.
These results are based on Anthropic’s own testing, so they should be viewed as company-reported benchmarks rather than an independent industry ranking.
Lower Costs Could Be the Bigger Story
One of the most notable changes with Fable 5.1 is its pricing.
Anthropic says typical token-based workloads should cost about 25% less than Fable 5. For highly agentic workloads, where the model repeatedly works with large amounts of previously processed context, savings can reach approximately 45%.
The company also cut the price of cache reads by 75%, bringing them to $0.25 per million tokens. Standard input and output pricing remains at $10 and $50 per million tokens, respectively.
The lower operating cost could make it easier for businesses and developers to use more capable AI systems for lengthy coding and research projects.
Fable 5.1 and Mythos 5.1 Are Built From the Same Model
Anthropic says Fable 5.1 and Mythos 5.1 are the same underlying model.
The difference is in their safeguards.
Fable 5.1 is generally available, while Mythos 5.1 is reserved for vetted users through trusted-access programs focused on cybersecurity and life sciences.
The company says Mythos 5.1 has stronger capabilities in areas where unrestricted access could create additional risks, which is why access is being limited.
AI Research Is Becoming a Bigger Focus
Anthropic is also highlighting scientific research as an important part of the new models.
In one example, the company says Mythos 5.1 designed high-affinity protein binders that were experimentally validated by outside organizations. Anthropic reported that some of the designs achieved binding affinities 10 times higher than the best designs submitted to certain protein-design competitions.
Fable 5.1 was also used to create a higher-resolution elevation map of part of Venus using data from NASA’s Magellan mission, according to Anthropic.
These examples illustrate where Anthropic believes advanced AI systems could eventually move beyond conventional chat and coding applications into scientific discovery.
Anthropic Is Expanding Its AI Safety Measures
The launch also includes changes to Anthropic’s safety systems.
The company says its latest cybersecurity safeguards produce about 60% fewer false positives than the previous safeguards on Fable 5. The company is also allowing Fable 5.1 to identify software vulnerabilities for defensive security work while continuing to restrict the development of exploits.
Anthropic says its biology safeguards also flag benign elementary biology and medical questions substantially less often than the safeguards introduced with Fable 5.
At the same time, advanced life sciences capabilities remain restricted to trusted users.
New Privacy Tools Target Businesses
Anthropic is also introducing Enterprise Frontier Safeguards, a system designed for organizations that want stronger control over their AI-related data.
Under the system, eligible customers can store activity data in cloud infrastructure controlled by the customer rather than Anthropic. The company says the system is designed to combine privacy protections with automated safety monitoring.
The safeguards are being developed with input from major organizations, including financial institutions and other companies that operate in heavily regulated industries.
Anthropic says Enterprise Frontier Safeguards will begin rolling out to enterprise customers in phases later this fall.
Claude Fable 5.1 Is Available Now
Claude Fable 5.1 is available today across Anthropic’s supported platforms, including Amazon Web Services, Google Cloud and Microsoft Azure.
Developers can access the model through Anthropic’s API using the claude-fable-5-1 model identifier.
Mythos 5.1 is more restricted, with Anthropic initially limiting availability to vetted cybersecurity and life sciences users. The company says it is working with the U.S. government to expand access to additional organizations.
What the Launch Means for the AI Race
The Fable 5.1 launch highlights a growing shift in the AI industry.
The competition is no longer only about producing better answers in a chatbot. Companies are increasingly focused on models that can write and debug software, conduct research, use computer tools and work through complicated assignments with less human supervision.
Anthropic’s decision to pair those improvements with lower token costs could make the development of long-running AI agents more practical for businesses.
The company is also betting that better safeguards and stronger privacy controls will help bring advanced AI into industries where data security and regulatory requirements have traditionally limited adoption.