ModernVBERT
ModernVBERTModernVBERT is a 250M-parameter vision-language encoder built for document retrieval rather than repurposed from a generative decoder. The paper walks the whole training pipeline - attention masking, image resolution, modality-alignment data regime, late-interaction contrastive objective - and the resulting ColModernVBERT checkpoint matches visual-document retrievers ten times its size while running query encoding on CPU. The suite also ships a bi-encoder variant, the aligned base model and the intermediate checkpoints.
From the same authors as ColPali - Teiletche, Mace, Conti, Loison, Viaud, Colombo and Faysse - and released under its own ModernVBERT Hub org rather than under vidore. The colpali-engine branch that loads it was not merged at the time of writing, which the card says outright. Verified 2026-09-11 via the ColModernVBERT model card, the arXiv preprint and the Hugging Face API.
Openness
5 high confidence- weights
- open(ungated MIT checkpoints on the ModernVBERT org - colmodernvbert, bimodernvbert, modernvbert-embed and the base model)
- data
- open(the paper states models, code and data are released, and the org publishes the NatCap modality-alignment corpus alongside the public document-retrieval sets it trains on)
- code
- open(the card releases the training codebase under MIT, and the paper releases intermediate checkpoints with it)
- checkpoints
- open(ModernVBERT/ablation_checkpoints, the intermediate and ablation runs behind the paper)
- license
- MIT(OSI)
An unusually complete release for a 2025 retriever: MIT over architecture, weights and the training codebase in one sentence of the card, the data published in the same org, and the intermediate checkpoints kept rather than discarded. The corpus is the softest of the four - it is a mixture of public sets plus the released NatCap rather than one named published corpus - but nothing in it is withheld.
- https://huggingface.co/ModernVBERT/colmodernvbert/raw/main/README.md recorded 2026-09-11
`license: mit` front matter and 'We release the ModernVBERT model architectures, model weights, and training codebase under the MIT license.'; lists ColModernVBERT, BiModernVBERT, ModernVBERT-embed and the base model.
- https://arxiv.org/abs/2510.01149 recorded 2026-09-11
'Models, code and data are available at this https URL.'
- https://huggingface.co/api/datasets?author=ModernVBERT&limit=50 recorded 2026-09-11
the org publishes ModernVBERT/natcap and ModernVBERT/toy-colpali-train-set as public datasets.
- https://huggingface.co/api/models?author=ModernVBERT&limit=50 recorded 2026-09-11
seven public repos including ModernVBERT/ablation_checkpoints; `gated: false` throughout.
Adoption
3 high confidence132,434 downloads in the trailing 30 days for ColModernVBERT, which is effectively the whole family - the bi-encoder, base and ablation repos draw single digits. The warehouse sweep recorded 248,469 over its own window; both readings sit inside the same band.
- https://huggingface.co/api/models/ModernVBERT/colmodernvbert recorded 2026-09-11
`downloads` (trailing 30 days) = 132434; `gated: false`; `private: false`.
Capability
3 high confidenceRung 3 is the narrow-scope reading, not the dated one: this is a 2025 model whose published numbers are credible and whose sub-1B efficiency is genuinely leading, but the scope is English-only at 250M and the absolute retrieval quality is a tier below the frontier on the same table.
- https://arxiv.org/html/2510.01149v1 recorded 2026-09-11
Table 3, ViDoRe Leaderboard, nDCG@5: ColModernVBERT (ours) 0.25B 81.2 / 56.0 / 68.6 at .032s CPU query latency; ColPali 2.92B 81.6 / 56.8 / 69.2; ColQwen2.5 3.75B 89.5 / 61.5 / 75.5; ColFlor 0.17B 68.8 / 43.0 / 55.9.
- https://arxiv.org/abs/2510.01149 recorded 2026-09-11
'a compact 250M-parameter vision-language encoder that outperforms recent models up to 10 times larger when fine-tuned on document retrieval tasks'.
Verified 2026-09-11