SenseNova-MARS: Empowering Multimodal Agentic Reasoning and Search via Reinforcement Learning
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908784746561536 |
|---|---|
| author | Chng, Yong Xien Hu, Tao Tong, Wenwen Li, Xueheng Chen, Jiandong Yu, Haojia Lu, Jiefan Guo, Hewei Deng, Hanming Xie, Chengjun Huang, Gao Lin, Dahua Lu, Lewei |
| author_facet | Chng, Yong Xien Hu, Tao Tong, Wenwen Li, Xueheng Chen, Jiandong Yu, Haojia Lu, Jiefan Guo, Hewei Deng, Hanming Xie, Chengjun Huang, Gao Lin, Dahua Lu, Lewei |
| contents | While Vision-Language Models (VLMs) can solve complex tasks through agentic reasoning, their capabilities remain largely constrained to text-oriented chain-of-thought or isolated tool invocation. They fail to exhibit the human-like proficiency required to seamlessly interleave dynamic tool manipulation with continuous reasoning, particularly in knowledge-intensive and visually complex scenarios that demand coordinated external tools such as search and image cropping. In this work, we introduce SenseNova-MARS, a novel Multimodal Agentic Reasoning and Search framework that empowers VLMs with interleaved visual reasoning and tool-use capabilities via reinforcement learning (RL). Specifically, SenseNova-MARS dynamically integrates the image search, text search, and image crop tools to tackle fine-grained and knowledge-intensive visual understanding challenges. In the RL stage, we propose the Batch-Normalized Group Sequence Policy Optimization (BN-GSPO) algorithm to improve the training stability and advance the model's ability to invoke tools and reason effectively. To comprehensively evaluate the agentic VLMs on complex visual tasks, we introduce the HR-MMSearch benchmark, the first search-oriented benchmark composed of high-resolution images with knowledge-intensive and search-driven questions. Experiments demonstrate that SenseNova-MARS achieves state-of-the-art performance on open-source search and fine-grained image understanding benchmarks. Specifically, on search-oriented benchmarks, SenseNova-MARS-32B scores 74.3 on MMSearch and 54.4 on HR-MMSearch, surpassing proprietary models such as Gemini-3-Pro and GPT-5.2. SenseNova-MARS represents a promising step toward agentic VLMs by providing effective and robust tool-use capabilities. To facilitate further research in this field, we will release all code, models, and datasets. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_24330 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | SenseNova-MARS: Empowering Multimodal Agentic Reasoning and Search via Reinforcement Learning Chng, Yong Xien Hu, Tao Tong, Wenwen Li, Xueheng Chen, Jiandong Yu, Haojia Lu, Jiefan Guo, Hewei Deng, Hanming Xie, Chengjun Huang, Gao Lin, Dahua Lu, Lewei Computer Vision and Pattern Recognition While Vision-Language Models (VLMs) can solve complex tasks through agentic reasoning, their capabilities remain largely constrained to text-oriented chain-of-thought or isolated tool invocation. They fail to exhibit the human-like proficiency required to seamlessly interleave dynamic tool manipulation with continuous reasoning, particularly in knowledge-intensive and visually complex scenarios that demand coordinated external tools such as search and image cropping. In this work, we introduce SenseNova-MARS, a novel Multimodal Agentic Reasoning and Search framework that empowers VLMs with interleaved visual reasoning and tool-use capabilities via reinforcement learning (RL). Specifically, SenseNova-MARS dynamically integrates the image search, text search, and image crop tools to tackle fine-grained and knowledge-intensive visual understanding challenges. In the RL stage, we propose the Batch-Normalized Group Sequence Policy Optimization (BN-GSPO) algorithm to improve the training stability and advance the model's ability to invoke tools and reason effectively. To comprehensively evaluate the agentic VLMs on complex visual tasks, we introduce the HR-MMSearch benchmark, the first search-oriented benchmark composed of high-resolution images with knowledge-intensive and search-driven questions. Experiments demonstrate that SenseNova-MARS achieves state-of-the-art performance on open-source search and fine-grained image understanding benchmarks. Specifically, on search-oriented benchmarks, SenseNova-MARS-32B scores 74.3 on MMSearch and 54.4 on HR-MMSearch, surpassing proprietary models such as Gemini-3-Pro and GPT-5.2. SenseNova-MARS represents a promising step toward agentic VLMs by providing effective and robust tool-use capabilities. To facilitate further research in this field, we will release all code, models, and datasets. |
| title | SenseNova-MARS: Empowering Multimodal Agentic Reasoning and Search via Reinforcement Learning |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2512.24330 |