Saved in:
| Main Authors: | , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2509.14033 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914046814453760 |
|---|---|
| author | Yin, Weijie Ye, Yongjie Shu, Fangxun Liao, Yue Kang, Zijian Dong, Hongyuan Yu, Haiyang Yang, Dingkang Wang, Jiacong Wang, Han Liu, Wenzhuo Liang, Xiao Yan, Shuicheng Feng, Chao |
| author_facet | Yin, Weijie Ye, Yongjie Shu, Fangxun Liao, Yue Kang, Zijian Dong, Hongyuan Yu, Haiyang Yang, Dingkang Wang, Jiacong Wang, Han Liu, Wenzhuo Liang, Xiao Yan, Shuicheng Feng, Chao |
| contents | We introduce SAIL-VL2, an open-suite vision-language foundation model (LVM) for comprehensive multimodal understanding and reasoning. As the successor to SAIL-VL, SAIL-VL2 achieves state-of-the-art performance at the 2B and 8B parameter scales across diverse image and video benchmarks, demonstrating strong capabilities from fine-grained perception to complex reasoning. Its effectiveness is driven by three core innovations. First, a large-scale data curation pipeline with scoring and filtering strategies enhances both quality and distribution across captioning, OCR, QA, and video data, improving training efficiency. Second, a progressive training framework begins with a powerful pre-trained vision encoder (SAIL-ViT), advances through multimodal pre-training, and culminates in a thinking-fusion SFT-RL hybrid paradigm that systematically strengthens model capabilities. Third, architectural advances extend beyond dense LLMs to efficient sparse Mixture-of-Experts (MoE) designs. With these contributions, SAIL-VL2 demonstrates competitive performance across 106 datasets and achieves state-of-the-art results on challenging reasoning benchmarks such as MMMU and MathVista. Furthermore, on the OpenCompass leaderboard, SAIL-VL2-2B ranks first among officially released open-source models under the 4B parameter scale, while serving as an efficient and extensible foundation for the open-source multimodal community. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_14033 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | SAIL-VL2 Technical Report Yin, Weijie Ye, Yongjie Shu, Fangxun Liao, Yue Kang, Zijian Dong, Hongyuan Yu, Haiyang Yang, Dingkang Wang, Jiacong Wang, Han Liu, Wenzhuo Liang, Xiao Yan, Shuicheng Feng, Chao Computer Vision and Pattern Recognition We introduce SAIL-VL2, an open-suite vision-language foundation model (LVM) for comprehensive multimodal understanding and reasoning. As the successor to SAIL-VL, SAIL-VL2 achieves state-of-the-art performance at the 2B and 8B parameter scales across diverse image and video benchmarks, demonstrating strong capabilities from fine-grained perception to complex reasoning. Its effectiveness is driven by three core innovations. First, a large-scale data curation pipeline with scoring and filtering strategies enhances both quality and distribution across captioning, OCR, QA, and video data, improving training efficiency. Second, a progressive training framework begins with a powerful pre-trained vision encoder (SAIL-ViT), advances through multimodal pre-training, and culminates in a thinking-fusion SFT-RL hybrid paradigm that systematically strengthens model capabilities. Third, architectural advances extend beyond dense LLMs to efficient sparse Mixture-of-Experts (MoE) designs. With these contributions, SAIL-VL2 demonstrates competitive performance across 106 datasets and achieves state-of-the-art results on challenging reasoning benchmarks such as MMMU and MathVista. Furthermore, on the OpenCompass leaderboard, SAIL-VL2-2B ranks first among officially released open-source models under the 4B parameter scale, while serving as an efficient and extensible foundation for the open-source multimodal community. |
| title | SAIL-VL2 Technical Report |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2509.14033 |