Anthropic has launched Claude Fable 5.1 and Claude Mythos 5.1, introducing faster performance, lower cache costs and fewer safeguard restrictions compared with the previous Fable 5 model.
The company said the updated models deliver improved benchmark scores while making AI usage more cost-efficient for developers. The new models arrive a few months after the launch of the original Fable 5. Anthropic has focused on reducing the cost of cache reads, improving model performance and adjusting safety measures following criticism around restrictions in the earlier version.
New Models Offer Lower Cache Costs
Anthropic has reduced the price of cache reads by 75%, bringing the cost down from the previous rate of $0.25 per million tokens. The company said normal pricing remains unchanged, with input tokens costing $10 per million and output tokens priced at $50 per million.
Cache reads are used when Claude reuses previously processed context, a feature that is particularly useful for coding agents and long-running tasks where the model repeatedly refers to the same instructions or repository information. According to Anthropic, the improvements make Fable 5.1 around 25% cheaper than Fable 5 for an average workload. For tasks involving heavier use of cached context, the savings could reach up to 45%.
Algoritha Security Launches ‘Make in India’ Cyber Lab for Educational Institutions
Improved Benchmarks and Reduced Restrictions
Anthropic said Claude Fable 5.1 has shown improvements across several internal tests. The model scored 55.8% on Terminal-Bench 4.0, compared with 42% for Fable 5, while CursorBench 3.2.0 performance increased from 70.5% to 73.4%.
The company also reported that Fable 5.1 more than doubled the previous model’s score on Anthropic’s Terminal-Bench-Science test, improving from 24.7% to 52.6%.
The updated model also includes changes to safeguards. Anthropic has relaxed some restrictions and allowed Fable 5.1 to identify software vulnerabilities, making it more suitable for defensive security work. The company said the changes could lead to about 60% fewer safeguard interventions during an average Claude Code session compared with the protections originally included with Fable 5.
Availability Across Major Cloud Platforms
Claude Fable 5.1 is now available through Claude, the Claude API, Amazon Web Services, Google Cloud and Microsoft Azure. Developers can access the model using the claude-fable-5-1 model ID.
Anthropic said the updated models combine performance improvements with reduced costs, allowing developers to use Claude more efficiently for coding and other long-running AI tasks.
Follow the Centre for Police Technology on LinkedIn to stay updated on the latest developments in policing, cybersecurity, digital forensics, investigations, fraud risk management, and technology-driven public safety.
https://www.linkedin.com/company/policetechnology/
