GLM-5.2 is live. #1 throughput on OpenRouter, pay-per-token on FriendliAI. Try it today ➜
wfwefw
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
wfwefw
Model Tree
Qwen/Qwen2.5-VL-3B-Instruct
Input Modalities
Output Modalities
Supported Functionality
QuIC-360 test set (5349 query-caption pairs):
| BLEU-4 | METEOR | ROUGE-L | CIDEr | SPICE |
|---|---|---|---|---|
| 0.0428 | 0.1137 | 0.2440 | 0.3428 | 0.1623 |
Three-seed aggregate results and exact split hashes are maintained in the CORA-360 repository.
bash
git clone --branch v2.0.0-eccv2026 https://github.com/wooseungw/CORA-360.gitcd CORA-360./reproduce.sh evaluate qwen2.5-vl-3b-vicreg /path/to/test.csv
CORA was evaluated on English query-focused captions from QuIC-360. Performance outside that domain, on non-ERP imagery, or in safety-critical settings is not established. Use of the base model remains subject to its upstream license.
bibtex
@inproceedings{woo2026cora,title={Overlap-Consistent View Decomposition for Adapting Vision--Language Models to 360-Degree Panoramas},author={Woo, Seungwoo and Jung, Daewon and Youm, Sekyoung},booktitle={European Conference on Computer Vision},year={2026}}