The work was developed with the investment research firm Citrini Research.
Over 64 hours, Claude generated 318,484 designs and selected 180 for testing, and Anthropic's team added 31 more with Claude's help.
Adaptyv Bio and Twist Bioscience, two contract research organizations, tested all 211 designs in the lab, and 17 met the criteria for a binder.
Anthropic has released all 211 designs and their data.
Anthropic cites 1.2 million or more scorpion stings a year and more than 3,250 deaths, mostly among children in regions with limited medical access.
Current antivenoms are made from antibodies in the blood of immunized horses, a process that can take months, and they usually work against only a few neurotoxins.
Anthropic says new treatments are costly to develop and unlikely to recoup their costs.
If binders can be designed in days and confirmed in weeks, it says, discovery for envenomation and other neglected diseases could speed up.
Outside experts Susana Vázquez Torres and Timothy P. Jenkins called the work a promising first step.
Vázquez Torres noted that binding is no longer the main bottleneck and that the harder task is making safe, affordable drugs.
The results have clear limits. Each of the 17 binders attaches to a single toxin: seven bind Cn2, four bind AaH II, four bind AaH I, and two bind LqhIII. None of the 91 designs aimed at all five toxins, or the designs aimed at toxin subgroups, met the criteria for more than one toxin. The two strongest binders, Cn2-B1 and Cn2-B2, bound Cn2 with affinities in the single-digit nanomolar range. Competition experiments suggest they bind at or near the site of LR, an antibody reported to neutralize Cn2 in mice. Anthropic stresses that it has not tested neutralization in cells or protection in animals, and that binding was measured only against lab-made toxins, so it does not claim to have produced an antivenom. Next steps would include confirming binding to toxin purified from venom, patch-clamp tests on sodium channels, and mouse studies.