%20--%3e%3cdefs%3e%3cstyle%3e%20.st0%20{%20fill:%20%23061b40;%20}%20.st1%20{%20fill:%20%23306af1;%20}%20.st2%20{%20fill:%20%235ce5cf;%20}%20%3c/style%3e%3c/defs%3e%3cg%3e%3cpath%20class='st0'%20d='M55,10.5h9v3h-9v9h12V7.5h-12v3ZM64,19.5h-6v-3h6v3Z'/%3e%3cpolygon%20class='st0'%20points='69%2016.5%2078%2016.5%2078%2019.5%2069%2019.5%2069%2022.5%2081%2022.5%2081%2013.5%2072%2013.5%2072%2010.5%2081%2010.5%2081%207.5%2069%207.5%2069%2016.5'/%3e%3cpolygon%20class='st0'%20points='95%2010.5%2095%207.5%2083%207.5%2083%2022.5%2095%2022.5%2095%2019.5%2086%2019.5%2086%2016.5%2095%2016.5%2095%2013.5%2086%2013.5%2086%2010.5%2095%2010.5'/%3e%3cpath%20class='st0'%20d='M40,1.5v21h11.6l1.4-1.4v-7.6h0c0,0-1.4-1.5-1.4-1.5l1.4-1.4V2.9l-1.4-1.4h-11.6ZM50,19.5h-7v-6h7v6ZM50,10.5h-7v-6h7v6Z'/%3e%3c/g%3e%3cpath%20class='st1'%20d='M23.1,24L14.7,4.8l-4.9,11.2h3.8l-1.8,4H3.3L12.1,0H2C.9,0,0,.9,0,2v20c0,1.1.9,2,2,2h21.1Z'/%3e%3cpath%20class='st2'%20d='M34,0h-16.8l10.6,24h6.2c1.1,0,2-.9,2-2V2C36,.9,35.1,0,34,0ZM32.5,20h-4V4h4v16Z'/%3e%3c/svg%3e)
Join Now
On September 1st local time, Anthropic announced the launch of Claude Fable 5.1 and Claude Mythos 5.1 models, calling them "the most advanced programming and knowledge work models in the world." Both models use the same base model, with differences in security protection levels: Fable 5.1 is available to all users, while Mythos 5.1 is only available to approved organizations in the fields of cybersecurity and life sciences through the Trusted Access Program.
Comprehensive Benchmarking Exceeds Previous Versions, with Lower Prices
In benchmark testing, Fable 5.1 outperformed its predecessor Fable 5, Opus 5, and OpenAI's GPT-5.6 Sol in multiple tests. In the Agent Science Research Benchmark Terminal-Bench-Science 0.1, Fable 5.1 scored 52.6%, compared to 24.7% for Fable 5, 29.0% for Opus 5, and 22.4% for GPT-5.6 Sol. In the Agent Programming Benchmark Terminal-Bench 4.0, Fable 5.1 scored 55.8%, while Mythos 5.1 achieved 60.9% under a more relaxed security setting. In knowledge work, Fable 5.1 scored 1853 points in the GDPval-AA v2 test, higher than Opus 5's 1824 points and Fable 5's 1723 points.
Price-wise, the cache read cost for Fable 5.1 has been reduced by 75%, dropping to $0.25 per million tokens. Under typical workloads, its cost is about 25% lower than Fable 5, and up to 45% lower for highly agent-based workloads. The base pricing remains unchanged: $10 per million input tokens and $50 per million output tokens.
In terms of security, Anthropic has improved the protective mechanisms of Fable 5.1 to reduce false positives. The new security measures in the cybersecurity field have reduced the number of false positive interceptions by about 60% compared to previous versions. Fable 5.1 can now be used to discover software vulnerabilities but will not be used to develop exploits for these vulnerabilities. In the biological field, Anthropic has partnered with the U.S. government to establish an access program that allows researchers to utilize Mythos 5.1's advanced biological capabilities, and plans to open registration for scientists soon.
Scientific Capabilities Realized, Compliance and Data Retention Upgraded
Anthropic showcased the achievements of Mythos 5.1 in multiple fields. In molecular design, Mythos 5.1 designed high-affinity binders using open-source protein design and folding tools, achieving binding affinities 10 times higher than the best entries in the Adaptyv Bio Protein Design Competition on three targets, with hit rates close to 50% on 12 targets, while the typical hit rate for protein design is currently 10% to 15%. In computational analysis, Fable 5.1 trained a neural network using radar images from NASA's Magellan mission over 30 years ago, creating a high-resolution elevation map covering one-third of Venus, with a resolution refined from 10 to 20 kilometers to 2 to 3 kilometers, and a 25% improvement in elevation accuracy. In computational biology, Mythos 5.1 increased the speed of seven open-source deep learning models by up to 2.5 times by writing custom GPU kernels and caching intermediate results, and is expected to reduce GPU costs by 30% to 60%.
In terms of data retention, Anthropic introduced Enterprise-Focused Security (EFS), combining zero data retention privacy protection with advanced security measures. Customer data is stored in cloud infrastructure fully controlled by the enterprise, rather than in Anthropic's systems. EFS will be rolled out in stages to enterprise customers this fall, and before that, eligible customers can use Fable 5.1 under the zero data retention policy.
In compliance, Anthropic signed the EU AI Act's practice guidelines on transparency of AI-generated content in July 2026. According to the law, model outputs released after August 2, 2026, must include an invisible watermark. Fable 5.1 has already added this invisible watermark to generated text, which is invisible to users without the detection API, does not affect output quality, and does not contain any information about the user or their conversation. Anthropic is launching the detection API in private preview, and qualified institutions such as regulatory agencies, law enforcement, media, and fact-checking organizations can use it according to EU laws.
Claude Fable 5.1 is now available on all platforms, including Amazon Web Services, Google Cloud, and Microsoft Azure, and developers can use it via the Claude API. Claude Mythos 5.1 is currently available only to approved U.S. organizations.