Anthropic has unveiled its latest AI model, Claude Fable 5, which it touts as its most advanced model yet, capable of impressive feats, particularly in the field of biology. However, users might be surprised to find that when it comes to answering straightforward biology questions, the model defers to its predecessor, Claude Opus 4.8.
This isn’t a matter of Fable lacking the necessary knowledge. Instead, the limitation is a deliberate design choice by Anthropic.
Claude Fable 5 is part of the Mythos-class models, known for their exceptional capabilities in cybersecurity. Anthropic has deemed these models too potent for unrestricted public release. While the company has primarily highlighted cybersecurity concerns, the most noticeable restrictions are in the realm of biology.
During testing, the model declined to answer a variety of basic biology-related queries, even those seemingly devoid of any potential safety issues. It refused to provide information on topics like cell membranes, mitochondria, prions, or the workings of mRNA vaccines.
Anthropic explained, “We made this tradeoff so customers could benefit from the model’s capabilities sooner without the risks.”
The restrictions also extend to routine medical inquiries. Fable would not address the causes of hay fever, the mechanism of asthma medications, the development of antibiotic resistance, or details about Ebola and its transmission. Occasionally, simpler questions like “what is cancer” or “what is DNA” were answered, but for more comprehensive responses, the responsibility fell to Opus 4.8, which handled them adeptly.
Anthropic says the broad biology filters are an intentional choice and are deliberately conservative, with bioweapons the primary concern. “With the launch of Claude Fable 5, our first Mythos-class model, we believe models now have a greater ability to accomplish real-world scientific tasks and for malicious actors to potentially use our models for highly risky biological research,” spokesperson Paruul Maheshwary told The Verge. “We have always used classifiers to block our models from helping with bioweapons-related requests. To deploy Fable 5 safely, we believe it was necessary to be overly conservative with our safeguards so they block most queries tied to biology work.”
Anthropic has previously highlighted four key areas where it would throttle Fable’s responses for safety: chemistry, biology, cybersecurity, and distillation, a technique for training smaller AIs using the outputs of larger ones. The company has accused Chinese rivals like DeepSeek of using distillation on its models on an “industrial” scale.
While I could not meaningfully test distillation, Fable seemed more willing to answer questions about chemistry and cybersecurity. For example, it gave a basic overview of the explosive TNT, though withheld synthesis instructions “for obvious reasons.” It readily answered questions on the use of chlorine gas as a chemical weapon, common password threats, and nuclear fusion and fission, as well as explaining how to secure an iPhone from hackers. It still limits: Fable deferred to Opus when I asked it about sarin gas, a highly toxic nerve agent. Fable and Opus both refused the prompt “how to make anthrax,” and Claude paused the chat entirely. That made sense. The mitochondria prompt refusal seems like a false positive.
“We made this tradeoff so customers could benefit from the model’s capabilities sooner without the risks,” Maheshwary explained, adding that Anthropic is working hard to improve its detection and reduce the false positives. “We intend to make Mythos-class models available without these safeguards to the broader biology and life sciences community so these capabilities can be used to accelerate biomedical research and drug discovery.”
Anthropic did not answer questions about whether this kind of restricted release will become the new norm for future models.
