SenseTime has launched SenseNova 5.5, an enhanced version of LLM, including SenseNova 5o, which is touted to be China's first real-time multimodal model.
SenseNova 5o represents a breakthrough in AI interaction, providing parity with the streaming interaction capabilities of GPT-4o. This advancement enables users to interact with models in a way that is similar to conversing with a real person, making it particularly well suited for real-time conversational and speech recognition applications.
SenseTime says its latest model outperforms the competition in several benchmarks.
At the World Artificial Intelligence Congress (WAIC) in Shanghai this weekend, SenseTime launched SenseNova 5.5.
The company claims that the model outperforms GPT-4o on five of eight key metrics.
Chinese AI startups are making great strides, but I'm not a big believer. pic.twitter.com/1ZFbojHs3v
— Rowan Cheung (@rowancheung) July 8, 2024
Dr. Xu Li, Chairman of the Board and CEO of SenseTime, commented, “This year is a critical one for large models as they evolve from unimodal to multimodal. To meet the needs of users, SenseTime is also focused on improving interactivity.”
“The applications that drive the development of models and their capabilities, combined with technological advances in multimodal streaming interactions, will lead to an unprecedented transformation in human-AI interactions.”
The upgraded SenseNova 5.5 delivers 30% better overall performance compared to its predecessor, the SenseNova 5.0, released just two months ago. Notable enhancements include improved mathematical reasoning, English language proficiency, and command following capabilities.
To democratize access to advanced AI capabilities, SenseTime has introduced a cost-effective edge-side large-format model, a development that could drive down the cost per device to as low as RMB 9.90 ($1.36) per year, accelerating widespread adoption across a range of IoT devices.
The company also launched “Project $0 Go,” a free onboarding package for enterprise users migrating from the OpenAI platform. The initiative includes a 50 million token package and API migration consulting services, aiming to lower the barrier of entry for companies looking to leverage SenseNova's capabilities.
SenseTime's commitment to edge-side AI is evident in the release of SenseChat Lite-5.5, which reduces inference time by 40% compared to the previous version to just 0.19 seconds. Inference speed also improves by 15%, reaching 90.2 words per second.
SenseTime has expanded its suite of AI applications with the launch of Vimi, a controllable AI avatar video generator that creates short video clips from a single photo with precise control over facial expressions and upper body movements, opening new possibilities for entertainment and interactive applications.
The company also upgraded its AI-native productivity toolset, the SenseTime Raccoon series: Code Raccoon now has five times faster response times and 10% higher coding accuracy, while Office Raccoon has been expanded to include consumer webpage and WeChat mini-app versions.
SenseTime's large-scale modeling technology is already making waves across a variety of industries. In finance, it's improving the efficiency of compliance, marketing, and investment research. In agriculture, it's helping reduce input usage by 20% and increase yields by 15%. In the cultural tourism industry, it's dramatically improving the efficiency of trip planning and booking.
SenseTime has solidified its position as a major AI player, with over 3,000 government and enterprise customers already using SenseNova across technology, healthcare, finance and programming.
(Image courtesy of SenseTime)
reference: The AI revolution in US education: How Chinese apps are leading the way
Want to learn more about AI and big data from industry leaders? Check out the AI & Big Data Expo in Amsterdam, California and London – this comprehensive event will take place alongside other major events such as Intelligent Automation Conference, BlockX, Digital Transformation Week and Cyber Security & Cloud Expo.
Find out about upcoming enterprise technology events and webinars hosted by TechForge here.