0.1.1

This page is for version 0.2.0 (latest) (default). For other versions, use one of these documentation indexes: - 0.2.0 (latest) (default): https://docs.nvidia.com/bionemo/inference-runtime/latest/llms.txt - 0.1.1: https://docs.nvidia.com/bionemo/inference-runtime/0.1.1/llms.txt - 0.1.0: https://docs.nvidia.com/bionemo/inference-runtime/0.1.0/llms.txt

Released 2026-10-02.

Fixed Issues

• Sequence validation. Protein, RNA, and DNA sequences with unsupported characters are now rejected instead of being silently treated as unknown residues. Ambiguous residues are normalized, and FASTA records that differ only by case are kept as one molecule. MSA rows with invalid characters are rejected.
• OpenFold3 confidence. pTM and ipTM now score only tokens that have a valid structural frame. Scores no longer drop valid ligand frames or include polymers with missing backbone atoms.
• Reproducible sampling. A seed set on a request is now applied when the model runs. Results no longer change based on how requests are batched. An explicit sampling seed still overrides the request seed.
• Safer file access. MSA and template paths stay inside the configured input directory, and written outputs stay inside the configured output directory.
• More reliable runs. Model artifacts are checked before use, installs are more stable, and inference performance and runtime behavior are improved.

The full public history for this release is on GitHub.