Home About us

SenseTime released Ririxin 5o, a real-time multi-modal streaming interaction benchmark against GPT-4o

IT Friends 2024/07/05 16:55

IT Friends, SenseTime released Ririxin 5o, real-time multi-modal streaming interaction benchmarking GPT-4o

Shanghai, July 5, 2024 – SenseTime, a strategic partner of the 2024 World Artificial Intelligence Conference and High-Level Conference on Global Governance of Artificial Intelligence (WAIC 2024), held the "Love Without Borders, Towards New Force" Artificial Intelligence Forum, and released the first WYSIWYG model in China, "RiRixin 5o", which benchmarks the interactive experience against GPT-4o and realizes a new AI interaction model.

By integrating cross-modal information, based on various forms such as sound, text, image and video, China's first WYSIWYG model "Ririxin 5O" brings a new AI interaction mode, that is, real-time streaming multimodal interaction. The scene also showed this innovative interaction mode for everyone——

At the beginning, the staff just said hello to "Ririxin 5O", and it automatically recognized the words on the chest strap worn by the staff's neck, judged that the scene was the venue of the World Artificial Intelligence Conference, and said that they could "study hard" in this place.

IT Friends, SenseTime released Ririxin 5o, real-time multi-modal streaming interaction benchmarking GPT-4o

Next, the staff took a cute puppy doll, and "Ririxin 5O" accurately described the puppy's appearance, expression and important wear - a white hat with the logo of SenseTime Technology, which was very popular for the home team.

On some more difficulty, just open any page of a book, "every day new 5o" can be automatically introduced, not a simple OCR recognition text, but recognition of graphics and text to give a good understanding of the summary, all this can be completed in an instant, truly real-time interaction.

IT Friends, SenseTime released Ririxin 5o, real-time multi-modal streaming interaction benchmarking GPT-4o

The staff also played a "painting skill" on the spot, and drew a stick figure rabbit casually, "Ri Ri Xin 5O" called the painting cute, and then the staff drew a smile expression, it captured the smile from this calm expression, the staff changed a stroke to draw the mouth big and added a tongue, "Ri Ri Xin 5O" immediately said that this expression was much happier after seeing it.

This interactive mode is especially suitable for applications such as real-time dialogue and speech recognition, with strong multi-task adaptability, which can naturally handle multiple tasks in the same model, and adaptively adjust behavior and output according to different contexts, and can realize the interactive experience of benchmarking GPT-4o is due to the overall improvement of the ability of the "Ririxin 5.5" basic model.

In just over two months, the new "Ririxin 5.5" system ushered in a number of upgrades, with an average increase of 30% in comprehensive performance compared with "Rixin 5.0", and the ability to follow instructions in mathematical reasoning, English ability and instruction following has been significantly enhanced, and the interactive effect and a number of core indicators have been benchmarked against GPT-4o.

IT Friends, SenseTime released Ririxin 5o, real-time multi-modal streaming interaction benchmarking GPT-4o

"Ririxin 5.5" adopts a hybrid end-to-cloud collaboration expert architecture to maximize cloud-edge-end collaboration and reduce inference costs, and the model training is based on more than 10TB tokens of high-quality training data, including a large number of synthetic thinking chain data, to improve inference thinking ability.

IT Friends, SenseTime released Ririxin 5o, real-time multi-modal streaming interaction benchmarking GPT-4o

In order to allow more enterprise users to access and use the powerful capabilities of the "RiRixin" large model system with a low threshold, SenseTime recently launched the "Large Model 0 Yuan Go" plan. All newly registered users of "Daily New" can get a number of free service packages involving calling, migration, training, etc. At the same time, SenseTime will also give away 50 million Tokens packages for free, and send exclusive moving consultants to help OpenAI users achieve zero-service cost migration.

This article is from Xinzhi self-media and does not represent the views and positions of Business Xinzhi.If there is any suspicion of infringement, please contact the administrator of the Business News Platform.Contact: system@shangyexinzhi.com