SenseTime, a prominent Chinese AI company, has recently unveiled its latest update, SenseNova 5.5 LLM, at the highly anticipated 2024 World Artificial Intelligence Conference & High-Level Meeting on Global AI Governance (WAIC 2024). This enhanced release introduces SenseNova 5o, an innovative real-time multimodal model that rivals OpenAI’s GPT-4o in terms of streaming interaction capabilities.
The upgraded SenseNova 5.5 boasts a significant 30% improvement in performance compared to its predecessor, SenseNova 5.0. Notable enhancements include enhanced mathematical reasoning, English proficiency, and command following abilities, elevating its interactivity and core indicators to match those of GPT-4o.
Dr. Xu Li, the Chairman of the Board and CEO of SenseTime, highlighted the pivotal nature of this release, stating, This year marks a crucial milestone for large models as they evolve from unimodal to multimodal functionalities. SenseTime remains committed to meeting user demands by focusing on enhancing interactivity.
In addition to the advanced model update, SenseTime has made strides in accessibility by introducing a cost-effective edge-side large model, reducing the device cost to as low as RMB 9.90 annually. This strategic move aims to facilitate wider deployment across various IoT devices, including smartphones, tablets, and in-vehicle computers.
Furthermore, SenseTime has expanded its application suite with the launch of Vimi, an AI avatar video generator capable of producing short video clips with precise facial expressions and upper body movements control from a single photo. Additionally, the SenseTime Raccoon Series has undergone upgrades to enhance coding precision and response speed in the Code Raccoon tool.
To minimize barriers for enterprise users, SenseTime initiated the Project $0 Go initiative, offering a complimentary onboarding bundle for new enterprise users transitioning from the OpenAI platform. The SenseNova Large Model has already been adopted by over 3,000 government and corporate clients across diverse sectors such as technology, healthcare, finance, and programming.
SenseTime remains committed to developing AI applications for vertical industries like finance, agriculture, cultural tourism, and healthcare, with the goal of enhancing productivity and cost-efficiency in these sectors.
Amidst these advancements, Kyutai, a French non-profit AI research lab, has unveiled Moshi, a real-time multimodal foundational AI model, presenting a voice-enabled AI assistant to challenge the functionalityctrd.
On the other hand, Anthropic’s Claude Sonnet 3.5 continues to disrupt the AI landscape by surpassing GPT-4o and securing the top spot in both the Coding Arena and Hard Prompts Arena.
The continuous evolution and innovation within the AI sector underscore the dynamic nature of technological advancement and competition globally.
