⚠️ This is a pretrained (base) model, not an instruction-tuned assistant. It is released as a reproducible reference point for data-centric research: to compare pretraining datasets, and as an initialization for your own SFT. It has not been chat- or preference-tuned and should not be expected to follow instructions reliably.

Model details


Scale	`large`
Parameters ($N$)	4B
Training tokens ($D$)	100B
Architecture	InternVL-2.5 (`InternVLChatModel`)
Vision encoder	InternViT-300M-448px
Language model	Qwen/Qwen2.5-3B
Precision	bfloat16
Training compute	~5,120 H100 hours
Training data	DCVLM-Baseline
Data mixture	10% captioning / 70% visual instruction / 15% text-only / 5% multimodal documents

Results

Core evaluation suite (33 tasks), compared against FineVision, the previous best open pretraining dataset, trained at the same scale with the same architecture and token budget.

Method	Gen	Know	OCR	Vision	MTL	Text	Core Avg
FineVision	59.0	70.7	58.9	39.1	45.1	51.2	54.2
DCVLM-Baseline (this model)	68.4	67.6	54.1	57.2	50.9	53.8	58.9

Categories: Gen general understanding · Know knowledge · OCR OCR & charts · Vision vision-centric · MTL multilingual · Text text-only.

Usage

The model uses custom modeling code, so trust_remote_code=True is required.

python

import torch
from transformers import AutoModel, AutoTokenizer

path = "mlfoundations/dcvlm-4b-model"
model = AutoModel.from_pretrained(
    path, torch_dtype=torch.bfloat16, trust_remote_code=True, low_cpu_mem_usage=True
).eval().cuda()
tokenizer = AutoTokenizer.from_pretrained(path, trust_remote_code=True, use_fast=False)

Image preprocessing follows the standard InternVL-2.5 dynamic-tiling recipe (448px tiles, max_dynamic_patch=12).

Related releases


Code	github.com/mlfoundations/dcvlm
Datasets	DCVLM-Baseline 6.25B · DCVLM-Baseline 200B · DCVLM-Balanced 200B
Models	1B · 2B · 4B · 8B

Citation

bibtex

@article{farina2026datacomp,
  title={DataComp-VLM: Improved Open Datasets for Vision-Language Models},
  author={Farina, Matteo and Udandarao, Vishaal and Nguyen, Thao and Kuzucu, Selim and B{\"o}ther, Maximilian and Hochlehnert, Andreas and Ghosh, Adhiraj and Nezhurina, Marianna and Roth, Karsten and Struber, Joschka and others},
  journal={arXiv preprint arXiv:2606.28551},
  year={2026}
}