BioTool: Biomedical Function-Calling Benchmark with 127 API Tools

BioTool is a large-scale benchmark and training corpus for the biomedical domain, pairing natural-language questions with correct tool calls. The corpus is drawn from 127 tools across three flagship public APIs: NCBI E-utilities, UniProt REST, and Ensembl REST. It was authored by gxx27 and last updated on 2026-05-08.

Use Cases

• Training biomedical language models for tool use based on the described question-tool call pairs.
• Benchmarking model performance on biomedical function-calling tasks based on the large-scale corpus.
• Fine-tuning models to interact with NCBI, UniProt, and Ensembl APIs based on the structured examples.

Strengths

• Large-scale corpus designed for a specific, complex task (biomedical function-calling).
• Covers 127 distinct tools from three major public biomedical APIs.
• Pairs natural-language questions with structured tool calls (function name and JSON arguments).

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and dataset size are unknown, which may limit suitability assessment.
• Last updated 2026-05-08 10:27:27; freshness should be verified for current API compatibility.

Provenance

Source

huggingface, author gxx27

Collection Method

Pairs questions with correct tool calls drawn from public APIs.

Time Range

null

Freshness

2026-05-08

Geography

null