Anthropic has introduced Claude Fable 5.1 and Claude Mythos 5.1, the world’s most advanced models for coding, knowledge work, and strong research capabilities. Fable 5.1 outperforms Fable 5, Opus 5, and OpenAI’s GPT 5.6 Sol across multiple benchmarks.Â
Higher Performance with Claude Fable 5.1Â
Claude Fable 5.1 sets new standards for coding, knowledge work, and long-running problem-solving tasks. When set to Low or Medium effort, Fable 5.1 delivers results similar or better than Fable 5’s at a lower cost.Â
Advanced Scientific Research Capabilities with Claude Mythos 5.1Â
Claude Fable 5.1 and Claude Mythos 5.1 show stronger capabilities in scientific research, including protein design, Venus’s mapping, and biology. Claude Mythos 5.1 achieved binding affinities 10 times higher than competitor benchmarks on 3 targets and a hit rate that reached nearly 50% across 12 targets. Alongside Anthropic’s new Enterprise Frontier Safeguards (EFS) system, gives customers privacy the same as zero data retention while being state-of-the-art at preventing adversarial use.Â
Stronger Cybersecurity and Vulnerability Research Â
The company has expanded what Fable 5.1 can do in vulnerability research. The model is used to discover software vulnerabilities. The defensive cybersecurity works without exploiting them. Alongside this, Anthropic says it has improved Fable 5.1’s safeguards, mainly for cybersecurity-related tasks. Claude Code users can see approximately 60% fewer false positives in cybersecurity, which could make the model helpful for security research and software development. Â
Anthropic has confirmed that Claude Mythos 5.1 is similar to Fable 5.1; however, both operate at different safeguard levels. Claude Mythos 5.1 will be available through trusted access programs. The Cyber Verification Program CVP provides access to Opus-and Sonnet-class models with reduced cyber safeguards for defensive security work. The Life Sciences Verification Program LSVP is designed for life sciences professionals who can use Claude Mythos 5.1, with safeguards for research and development.Â
Transparency and Compliance with EU AI Act Â
Anthropic has also signed the EU AI Act’s Code of Practice on Transparency of AI-Generated Content. This adds a watermark to the text generated by Claude and to the outputs of models released after August 2, 2026. This watermark is invisible to anyone without the detection API. Anthropic is thus rolling out a detection API in private preview. It is currently available to eligible organizations such as regulators, media, law enforcement, independent researchers, educational organizations, and EU civil society groups as required under EU law.Â
Pricing and AvailabilityÂ
Another significant update is pricing. Anthropic says they have reduced the price of Fable 5.1 cache reads, in which the model reuses context it has already processed, and usage is billed per token. Cache can now be read at 75% lower cost, or $0.25 per million tokens. Fable 5.1 is approximately 25% cheaper than Fable 5 for typical workloads. For complex coding and agentic tasks, the savings are up to around 45%.Â
Fable 5.1’s pricing is the same as Fable 5. It costs $10 per million input tokens and $50 per million output tokens.Â
Claude Fable 5.1 is available on all platforms, including Amazon Web Services, Google Cloud, and Microsoft Azure. This gives developers and businesses access to improved coding, security, and agentic capabilities. Developers can start with claude-fable-5.1 on the Claude API.Â
Claude Mythos 5.1 is available to cyber defenders and life scientists. It is only available to a set of organizations in the US. Overall, the new release represents Anthropic’s push towards a model that can handle complex, autonomous workflows while delivering state-of-the-art performance, safeguards, and lower operating costs. Â
Head over to our site to stay informed with all the latest news around the tech world.
Read More : Anthropic Launches Claude Fable 5 for Public Use with Claude Mythos 5