Anthropic’s Newly Launched Claude Fable 5.1 Costs Less and Codes Better

Anthropic has launched Claude Fable 5.1 and Claude Mythos 5.1 for coding, research, and complex knowledge work. Claude Fable 5.1 and Claude Mythos 5.1 use the same underlying model, but Mythos 5.1 has different safeguards for vetted cybersecurity and life sciences users. According to Anthropic, the new models can handle longer, more complex tasks while reducing costs for users. 

“They’re the world’s most advanced models for coding and knowledge work, and their research capabilities offer an early glimpse of how AI models will contribute to scientific progress,” Anthropic said in a statement. 

How Does Claude Fable 5.1 Reduce Costs for Users? 

Claude Fable 5.1 is designed to reduce token costs and false-positive interventions from its safeguards. For typical workloads, Anthropic said Claude Fable 5.1’s  pricing will be about 25% lower than that of Fable 5. The company charges usage by tokens. 

For highly agentic tasks where AI works through multiple steps using tools, Anthropic said users can save as much as 45%. The AI developer attributed the lower cost to a 75% drop in cache reads, which allow Fable 5.1 to reuse previously processed information. 

Anthropic said the new AI model will cost $10 per million input tokens and $50 per million output tokens. Cache reads will cost $0.25 per million tokens. Anthropic noted that Fable 5.1’s performance surpassed that of Fable 5 on various tests. 

For instance, Fable 5.1 scored 73.4% compared to Fable 5’s 70.5% on the CursorBench 3.2 test. Both Fable 5.1 and Mythos 5.1 a;so performed better in AI capability assessments. For instance,  Claude Fable 5.1 scored  higher in agentic coding than both Fable 5 and Opus 5. 

How Does Fable 5.1. Help With Cybersecurity? 

In cybersecurity, Claude Fable 5.1 can help identify software vulnerabilities as part of defensive security work. However, the model’s ability to perform tasks like penetration testing and exploit generation is restricted. 

Anthropic noted that Fable 5.1’s upgraded cybersecurity safeguards produce about 60% fewer false-positive interventions than the previous safeguards. The company has partnered with the U.S. government to set up an access program for professionals using Mythos 5.1 for research and development. 

Anthropic has also introduced Enterprise Frontier Safeguards for corporate customers. This system is designed to provide privacy while safeguarding AI models against misuse. Under this system, Anthropic will store data in cloud infrastructure controlled by the customer. The AI firm plans to roll out the system to enterprise customers later this year. 

How Can Users Access Anthropic’s New Models?

Anthropic has made Claude Fable 5.1’s availability less restrictive. Users can access the new AI model immediately via cloud platforms like Microsoft Azure and Google Cloud or Anthropic’s API. Access to Mythos 5.1 is restricted to registered partners working in life sciences and cybersecurity.