SenseTime Releases Open-Source SenseNova-MARS 8B/32B: Autonomously Tackles Complex Tasks, Outperforms Gemini-3-Pro

4 day ago / Read about 0 minute

Author：小编

Today, SenseTime made an official announcement regarding the open-sourcing of its multimodal autonomous reasoning model, SenseNova-MARS. This model comes in two variants: 8B and 32B. In the core benchmark tests focused on multimodal search and reasoning, the model scored an impressive 69.74, outperforming both Gemini-3-Pro, which scored 69.06, and GPT-5.2, which scored 67.64. SenseNova-MARS stands out as the first Agentic VLM model that facilitates a seamless integration of dynamic visual reasoning with text-image search capabilities. This integration empowers the model to autonomously plan steps, invoke necessary tools, and effectively manage complex tasks, showcasing its 'executive capabilities.'

Previous page：OpenAI Steps Up IPO Pace, Targeting Q4 2026 and Vy...

Next page：High Precision, Small Footprint, and Free for Comm...

Return to List

Hot Reading

2 day ago

HomeBoost’s app will show you where to save on your utility bills

2 day ago

Nvidia CEO pushes back against report that his company’s $100B OpenAI investment has stalled

2 day ago

8 Alexa Settings Everyone Should Disable on New Amazon Echos

2 day ago

I Tested the New AirTag and Found That Apple More Than Doubled Its Range