Anthropic says Claude Fable 5.1 ‘sets a new standard for coding, knowledge work, and long-running problem-solving tasks’

Anthropic has cut the ribbon on its Claude Fable 5.1 AI model, bringing performance enhancements and security improvements – along with new cost efficiency boosts.

According to the firm, Fable 5.1 “sets a new standard for coding, knowledge work, and long-running problem-solving tasks”.

Internal benchmark tests showed the new model had significantly higher performance capabilities than Fable 5 across four key metrics: agentic scientific research, agentic terminal coding, multidisciplinary reasoning, and agentic coding.

Anthropic also claimed that Fable 5.1 is roughly 25% cheaper than its predecessor for “typical workloads”, providing usage is billed on a per-token basis. Anthropic said this is due to price reductions in cache reads, which involve the model reading inputs that have already been processed and stored.

“For highly agentic work, the savings will often be much larger—up to approximately 45%,” the company said in a blog post.

These improvements come amidst rising concerns about AI-related costs, which are largely down to how the technology is now being used. The shift to agentic AI is a key factor here, mainly as agents typically consume more tokens compared to traditional chatbot-based interactions.

Research from Signal65 found that agentic AI workloads consume anywhere up to fifteen-times more tokens compared to chatbots.

New safeguards for Claude Fable 5.1

Anthropic also revealed a series of new safeguards for Fable 5.1. This includes new Enterprise Frontier Safeguards (EFS), a system that gives customers “complete privacy” through a zero data retention policy.

EFS works by storing data in cloud infrastructure “controlled entirely by the customer”, the company said and is expected to roll out later this year for enterprise customers.

New cybersecurity-related safeguards have also been introduced, with particular focus on reducing false positives. Anthropic said changes rolled out with Fable 5.1 deliver a 60% reduction in false positives than before.

“In part, this is because Fable 5.1 can now be used to discover software vulnerabilities – though not to develop exploits for them,” the company explained.

Prior to Fable 5.1’s release, Anthropic said it performed “extensive stress-testing” of safeguards to prevent potential jailbreaks and misuse.

“In addition to our own dynamic evaluations of their robustness, we commissioned external testing from two organisations, along with automated testing by Gray Swan,” the company said.

“As with Fable 5 and Opus 5, we have not found evidence of a critical-severity jailbreak for these safeguards.”

Fable 5.1 is available now for customers through selected cloud platforms, or through the Anthropic API.

Mythos 5.1, meanwhile, will only be available to registered partners conducting cybersecurity of life science research. This continues the company’s strategy of locking down access to selected partners due to security-related concerns.

FOLLOW US ON SOCIAL MEDIA

Follow ITPro on Google News and add us as a preferred source to keep tabs on all our latest news, analysis, views, and reviews.

You can also follow ITPro on LinkedIn, X, Facebook, and BlueSky.