GLiNER is a Named Entity Recognition (NER) model capable of identifying any entity type using a bidirectional transformer encoders (BERT-like). It provides a practical alternative to traditional NER models, which are limited to predefined entities, and Large Language Models (LLMs) that, despite their flexibility, are costly and large for resource-constrained scenarios.

GLiNER-biomed, developed in collaboration with DS4DH from the University of Geneva, introduces a specialized suite of efficient open biomedical NER models based on the GLiNER framework. GLiNER-biomed leverages synthetic annotations distilled from large generative biomedical language models to achieve state-of-the-art zero-shot and few-shot performance in biomedical entity recognition tasks.

How to Use

Install the official GLiNER library with pip:

pip install gliner -U

After installing the GLiNER library, you can easily load a GLiNER-biomed model and perform named entity recognition:

python

from gliner import GLiNER

model = GLiNER.from_pretrained("Ihor/gliner-biomed-small-v1.0")

text = """
The patient, a 45-year-old male, was diagnosed with type 2 diabetes mellitus and hypertension.
He was prescribed Metformin 500mg twice daily and Lisinopril 10mg once daily. 
A recent lab test showed elevated HbA1c levels at 8.2%.
"""

labels = ["Disease", "Drug", "Drug dosage", "Drug frequency", "Lab test", "Lab test value", "Demographic information"]

entities = model.predict_entities(text, labels, threshold=0.5)

for entity in entities:
    print(entity["text"], "=>", entity["label"])

Expected output:

45-year-old male => Demographic information
type 2 diabetes mellitus => Disease
hypertension => Disease
Metformin => Drug
500mg => Drug dosage
twice daily => Drug frequency
Lisinopril => Drug
10mg => Drug dosage
once daily => Drug frequency
HbA1c levels => Lab test
8.2% => Lab test value

Benchmarking

We examined our models on 8 complex real-world datasets and compared them with other GLiNER models.

Model	F1-score	Macro mean F1	Macro median F1	Weighted F1
Large models
NuNER Zero	40.87	21.79	13.94	33.67
NuNER Zero span	40.26	22.51	14.27	32.52
GLiNER bio v0.1	42.34	27.10	24.44	38.32
GLiNER bio v0.2	38.66	25.36	17.02	32.42
GLiNER v1.0	47.77	29.60	21.13	40.78
GLiNER v2.0	37.38	21.42	15.44	33.11
GLiNER v2.1	48.04	29.75	28.20	43.43
GLiNER news v2.1	48.99	31.79	33.77	45.13
GLiNER v2.5	53.81	35.22	35.65	51.57
GLiNER-biomed	59.77	40.67	42.65	58.40
GLiNER-biomed-bi	54.90	35.78	31.66	50.46
Base models
GLiNER v1.0	41.61	24.98	10.27	31.59
GLiNER v2.0	34.33	24.48	22.01	30.58
GLiNER v2.1	40.25	25.26	14.41	32.64
GLiNER news v2.1	41.59	27.16	17.74	34.44
GLiNER v2.5	46.49	30.93	25.26	44.68
GLiNER-biomed	54.37	36.20	41.61	53.05
GLiNER-biomed-bi	58.31	35.22	32.39	54.91
Small models
GLiNER v1.0	40.99	22.81	7.86	31.15
GLiNER v2.0	33.55	21.12	15.76	28.78
GLiNER v2.1	38.45	23.25	10.92	30.67
GLiNER news v2.1	39.15	24.96	14.48	33.10
GLiNER v2.5	38.21	28.53	18.01	36.88
GLiNER-biomed	52.53	34.49	38.17	50.87
GLiNER-biomed-bi	56.93	33.88	33.61	53.12

Join Our Discord

Connect with our community on Discord for news, support, and discussion about our models. Join https://discord.gg/dkyeAgs9DG.

Citation

This work

If you use GLiNER-biomed models in your work, please cite:

bibtex

@misc{yazdani2025glinerbiomedsuiteefficientmodels,
      title={GLiNER-BioMed: A Suite of Efficient Models for Open Biomedical Named Entity Recognition}, 
      author={Anthony Yazdani and Ihor Stepanov and Douglas Teodoro},
      year={2025},
      eprint={2504.00676},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2504.00676}, 
}

Previous work

bibtex

@misc{zaratiana2023gliner,
      title={GLiNER: Generalist Model for Named Entity Recognition using Bidirectional Transformer},
      author={Urchade Zaratiana and Nadi Tomeh and Pierre Holat and Thierry Charnois},
      year={2023},
      eprint={2311.08526},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}

bibtex

@misc{stepanov2024gliner,
      title={GLiNER multi-task: Generalist Lightweight Model for Various Information Extraction Tasks},
      author={Ihor Stepanov and Mykhailo Shtopko},
      year={2024},
      eprint={2406.12925},
      archivePrefix={arXiv},
      primaryClass={id='cs.LG' full_name='Machine Learning' is_active=True alt_name=None in_archive='cs' is_general=False description='Papers on all aspects of machine learning research (supervised, unsupervised, reinforcement learning, bandit problems, and so on) including also robustness, explanation, fairness, and methodology. cs.LG is also an appropriate primary category for applications of machine learning methods.'}
}

Source Model Repo

This model is derived from Ihor/gliner-biomed-small-v1.0. See the upstream repository for the original safetensors weights, training data, and the full upstream model card.

ONNX Weights

ONNX weights added by SAS — converted from the upstream safetensors checkpoint.

File in this repo: model.onnx.

Using this Model with the SAS GLiText API

This repo is consumed by the SAS GLiText product. To download it onto a SAS GLiText server:

POST /v1/models/download?name=small-biomed

To download and load into memory in one step:

PUT /v1/models?name=small-biomed

Security Scan

Scanned with modelaudit v0.2.40 on 2026-04-27. 29/29 checks passed. Full results.

File	Size	SHA-256
`model.onnx`	611.3 MB	`3802e22c6358e1c9…`

rpeel/glitext-small-biomed

GLiNER-BioMed

How to Use

Benchmarking

Join Our Discord

Citation

This work

Previous work

Source Model Repo

ONNX Weights

Using this Model with the SAS GLiText API

Security Scan

rpeel/
glitext-small-biomed