rewire.it
Paper-reported evidence

Benchmarking cell type and gene set annotation by large language models with AnnDictionary

2025 · Peer-reviewed · PMC archival version PMC12569162.1

Paper-reported results. The values below come from this source, not an independent rewire.it run.

Open primary paper →DOI: 10.1038/s41467-025-64511-x

Primary full text verified using Europe PMC XML; venue: Nature Communications; PMC ID: PMC12569162.

2 verified numerical rows · source checked 2026-09-15

Paper-reportedIndependent paper evaluation

Claude 3.5 Sonnet

Single-cell type label annotation

Reported score
84.0%± 0.7
Metric
Binary agreement with manual annotation
Dataset / split
AnnDictionary cell-type annotation test

LLM cell-type labels; binary agreement with manual annotations across five replicates.

Table 1, Claude 3.5 Sonnet row, Cell Type Binary Cells columnVerify at source →
Paper-reportedIndependent paper evaluation

GPT-4o

Single-cell type label annotation

Reported score
80.9%± 0.7
Metric
Binary agreement with manual annotation
Dataset / split
AnnDictionary cell-type annotation test

LLM cell-type labels; binary agreement with manual annotations across five replicates.

Table 1, GPT-4o row, Cell Type Binary Cells columnVerify at source →