Anthropic open-sources the code Claude wrote to make 30 protein models about 4x faster, and to fold a 70,000-token ribosome on one GPU node

Anthropic open-sourced code that its Claude model wrote to optimize more than 30 open-source protein, genomics and sequence-modeling models, reporting an average speedup of about 4x with small precision loss and about 1.6x with identical outputs. Two Anthropic staff with biology backgrounds and no kernel-engineering experience supervised Claude for about four weeks; the resulting FlashPairformer kernels run triangle attention 2.7 to 2.9x faster and triangle multiplication 1.7 to 3.2x faster than the field standard, and a 70,000-token bacterial ribosome ran on a single B300 node. Anthropic says a binder design campaign that previously took thousands of H100 hours cost about $150 in GPU and tokens, though the speedups are Anthropic's own measurements and have not yet been replicated by independent labs.

Research post and GitHub repo, surfaced on Anthropic's site this week. Two Anthropic staff with biology backgrounds and no kernel-engineering experience supervised Claude for about four weeks. Claude optimized more than 30 open-source models for structure prediction, protein design, genomics and sequence modeling. Average speedup about 4x with small precision loss, about 1.6x with identical outputs. New FlashPairformer kernels run triangle attention 2.7 to 2.9x faster and triangle multiplication 1.7 to 3.2x faster than the field standard. A 'big mode' lets a single NVIDIA node predict structures over 10,000 tokens, which was not possible before, and a 70,000-token bacterial ribosome ran on one B300 node. Anthropic says a binder design campaign that used to take thousands of H100 hours cost about $150 in GPU and tokens. All the code is in the repo. Why it matters: it is a concrete, checkable example of an agent doing weeks of specialist engineering, and the kernels are free to use. What to watch: the speedups are Anthropic's measurements. Independent labs have not replicated them yet.