Skip to content
Sign In Subscribe
Artificial Intelligence — Technology

Anthropic Launches Public Version Of Fable 5 With Safety Restriction

Anthropic said Fable 5 delivers a substantial improvement in reasoning and complex problem-solving performance.

,   and  TIPP Staff
3 min read
Anthropic releases Claude Fable 5. Pic via(@BusinessInsider)

Anthropic has publicly released Claude Fable 5, a new artificial intelligence model built on its advanced Mythos architecture but equipped with extensive safeguards designed to limit potentially dangerous uses.

The launch follows a tightly controlled rollout of the original Mythos Preview system earlier this year.

💡
According to reports, the preview model attracted significant attention from governments, regulators, and financial institutions because of its advanced analytical capabilities and its reported ability to identify software vulnerabilities at a highly sophisticated level.

Anthropic said Fable 5 delivers a substantial improvement in reasoning and complex problem-solving performance.

The company claims it is the first publicly available AI model to exceed 90% on the Hex benchmark, a test designed to measure long-horizon analytical and decision-making tasks.

💡
To address safety concerns, Anthropic has implemented automated safeguards that redirect requests involving sensitive subjects such as cybersecurity and biology to a more restricted model.

The company said these protections are intended to reduce the risk of misuse while still allowing broad public access to advanced AI capabilities.

At the same time, Anthropic has limited access to the full Mythos 5 model, making it available only through a specialized cybersecurity initiative involving select infrastructure providers and U.S. government cyber defense organizations.

The approach reflects growing efforts within the AI industry to balance innovation with security and risk management.

Related Tweet:

Also Read:

Why Anthropic Says Its Claude AI Is Learning To Reflect Like Humans
Anthropic, the AI company behind Claude, told Axios that its newest models are beginning to show “introspective awareness” — the ability to reflect on how they think. Anthropic researcher tells @axios it’s most advanced models show clear signs of human-like thinking about how they think… https://t.co/es4Bkdr7xW — Jim
tippinsightsTIPP Staff
Complaints Grow Over Changes To Anthropic’s Claude
Anthropic is facing backlash from power users who say its Claude AI model has declined in performance, sparking concerns about access to advanced tools, as reported by Axios. Users across platforms like GitHub and Reddit claim Claude now produces less accurate and nuanced responses, especially for complex tasks. Some have
tippinsightsTIPP Staff

Comments

Related

Latest