Portuguese 
English 
Spanish 
3 min of reading 
0 comments 
New model from Anthropic promises high performance in programming, science, and reasoning, but limits requests related to cybersecurity, biology, and chemistry.
Anthropic, the American company responsible for the chatbot Claude, announced on June 9, 2026 the launch of Claude Fable 5, its new artificial intelligence open to the public.
The model was described by the company as the most powerful AI ever made available by the company for general users.
The novelty is part of the Mythos class, a generation of models presented by Anthropic in April 2026.
-
British engineer turns a childhood dream inspired by Star Wars into a 1.9-ton mechanical spider: a machine with six hydraulic legs, a turbo diesel engine, and a Linux brain that can be piloted from inside like a giant steel creature.
-
Brazil filled the Northeast with wind turbines but now faces the side effect of 705 giant blades reaching the end of their lifespan by 2032 just in Rio Grande do Norte, while the world projects 43 million tons of waste by 2050.
-
Brazil has filled the Northeast with wind turbines, but now it faces the side effect of 705 giant blades reaching the end of their useful life by 2032 just in Rio Grande do Norte, while the world projects 43 million tons of waste by 2050.
-
Three astronauts launch towards the Chinese space station Tiangong, and for the first time, one of them will spend an entire year in orbit on a mission that includes the first astronaut from Hong Kong and prepares for a manned landing on the Moon by 2030.
This new family was positioned above the Opus line, which until then represented the company’s most advanced level.
According to Anthropic, the Claude Fable 5 achieves cutting-edge performance in software engineering, data analysis, scientific research, computer vision, and complex reasoning.
Technical investigation reveals limits in risk areas
The public launch came with restrictions on topics considered sensitive by the company.
Requests related to cybersecurity, biology, chemistry, and techniques for extracting knowledge from AI models may be automatically redirected.
In these situations, the conversation may be transferred to Claude Opus 4.8, a less powerful version of the system.
According to Anthropic, the measure was adopted because the Mythos class models have reached a level of capability that requires greater control.
The main concern is the malicious use of technology to discover and exploit software vulnerabilities.
This type of resource could facilitate more sophisticated cyberattacks, according to an assessment released by the company.
Impacts for Claude Fable 5 users
Anthropic states that more than 95% of usage sessions should not be affected by the restrictions.
Even so, the company acknowledges that some legitimate requests may end up blocked due to excessive caution.
This decision shows an attempt to balance public access, technological advancement, and risk reduction.
The stated goal is to allow broad use of Claude Fable 5, without fully releasing capabilities considered dangerous.
Claude Mythos 5 will have controlled access
At the same time, Anthropic launched Claude Mythos 5, a version of the same model with some safeguards removed.
Initial access will be restricted to selected partners linked to cyber defense and the protection of critical infrastructures.
This use will occur through Project Glasswing, a program developed in collaboration with the United States government.
The company stated that it intends to gradually expand access through a program aimed at users considered trustworthy.
According to Anthropic, Mythos 5 brings together the most advanced cybersecurity capabilities among its current models.
The technology can identify and exploit software flaws with unprecedented speed and precision, according to the company.
Scientific applications draw attention
Besides digital security, Anthropic highlighted the model’s advancements in scientific research.
The Mythos 5 is said to have accelerated stages of drug development and assisted in protein design.
The technology also generated new hypotheses in molecular biology, later evaluated in experiments by researchers.
Another point mentioned by the company involves research in genomics conducted largely autonomously for over a week.
During this period, millions of cells from different animal species were analyzed by the system.
Anthropic also claims that the model created its own machine learning structures to interpret the data.
What changes in the artificial intelligence race?
The launch of Claude Fable 5 reinforces the global race for more advanced artificial intelligence models.
The new release also shows how companies in the sector are trying to combine innovation, security, and access control.
While the public receives a powerful version of Claude, sensitive areas are now subject to stricter barriers.
The challenge for Anthropic will be to prove that technology can advance without increasing risks in cybersecurity, science, and misuse of AI.
Do you believe that advanced artificial intelligence models should be released without restrictions or need limits on sensitive topics? Share your opinion!
Be the first to react!