GLM-5.2 is live. #1 throughput on OpenRouter, pay-per-token on FriendliAI. Try it today ➜
ch-min
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Run this model inference with full control and performance in your environment.
Model Details
Model Provider
ch-min
Model Tree
Qwen/Qwen2.5-VL-3B-Instruct
Input Modalities
Output Modalities
Supported Functionality
This checkpoint inherits the architecture of Qwen/Qwen2.5-VL-3B-Instruct.
You can load it using the standard transformers interface:
python
from transformers import Qwen2_5_VLForConditionalGeneration, AutoProcessorimport torchmodel_id = "ch-min/Qwen2.5-VL-3B-Instruct-data_scale_exp_80k-20251114_120221" # Replace with the actual repository IDmodel = Qwen2_5_VLForConditionalGeneration.from_pretrained(model_id, torch_dtype="auto", device_map="auto")processor = AutoProcessor.from_pretrained(model_id)# For inference, follow the standard Qwen2.5-VL usage pattern
If you use this checkpoint, please cite both our paper and the original Qwen2.5-VL-3B paper.
Our paper (this checkpoint family):
bibtex
@article{min2026whyfarlooksup,title = {Why Far Looks Up: Probing Spatial Representation in Vision-Language Models},author = {Min, Cheolhong and Jung, Jaeyun and Lee, Daeun and Jeon, Hyeonseong andSu, Yu and Tremblay, Jonathan and Song, Chan Hee and Park, Jaesik},journal = {arXiv preprint arXiv:2605.30161},year = {2026},}
Original Qwen2.5-VL-3B (Qwen2.5-VL Technical Report):
bibtex
@article{bai2025qwen25vl,title = {Qwen2.5-VL Technical Report},author = {Bai, Shuai and Chen, Keqin and Liu, Xuejing and Wang, Jialin and others},journal = {arXiv preprint arXiv:2502.13923},year = {2025},}