The automotive robotics brand that has faced complaints about its complexity is determined to step back into the spotlight with the ambition of becoming the "strongest intelligent driving solution on Earth."
On August 28, Chengdu hosted the end-to-end AI Intelligent Driving Launch Conference themed "Evolution: AI Reshapes Intelligent Driving."
At the conference, the brand unveiled its most significant update yet: the new V2.0 version software, which will debut on the Jiyue 07.
This V2.0 version software introduces a comprehensive suite of enhancements, covering essential modules such as end-to-end large model intelligent driving, large model intelligent cabin, and intelligent vehicle connectivity, totaling 947 new features and upgrades.
The pre-sale price for the Jiyue 07 was announced on the same day, starting at 215,900 yuan, with plans for its official launch on September 10.
"You can’t do intelligent driving well without 50 billion."
With this update, the Jiyue PPA intelligent driving has been upgraded to ASD (Apollo Self Driving), marking Jiyue's transition into the era of end-to-end large model advanced intelligent driving.
Jiyue's CEO, Xia Yiping, stated, “This is the first time L4 autonomous driving technology has been applied to consumer vehicles! It’s the world's first major autonomous driving model to support fully unmanned driving applications—thanks to the deep integration of Baidu ADFM with Jiyue.”
He believes that Baidu's continuous research and development in AI and autonomous driving for over a decade, alongside an investment exceeding 50 billion, has been crucial for Jiyue’s sustained leadership.
He also remarked, “Without 50 billion, you can’t achieve good intelligent driving; any intelligent driving made with just a few billion is merely a ‘road killer.’”
Although there are still debates within the industry regarding the performance of pure visual intelligent driving solutions, Jiyue remains steadfastly committed to it, confident that the introduction of this end-to-end large model elevates its perception capabilities even above mainstream lidar technology.
Reports indicate that Jiyue's pure visual advanced intelligent driving features a perception range of up to 250 meters, with the ability to identify a water bottle within 200 meters. Moreover, the ASD system includes 106 new functions and upgrades in areas such as navigational assistance, automated parking, advanced lane keeping, and active safety—completely evolving driving, safety, and parking capabilities.
Currently, Apollo's L4 autonomous driving has surpassed 100 million kilometers of mileage, establishing a high-performance training cluster exceeding 5.5 EFLOPS. By leveraging AI-powered computing, Jiyue can detect collision risks 5,000 times per millisecond and perform 5,000 distance calculations, actively searching for the best driving trajectory. By the end of 2024, users will be able to use ASD wherever Baidu navigation maps are available, achieving nationwide accessibility.
The APA automated parking assistance not only supports various complex parking scenarios such as vertical, horizontal, spatial, inverted, and dead-end parking but has also overcome challenges like low light and extremely tight parking spots, doubling the success rate for parking.
As a result, Jiyue has initiated a free trial of pure visual advanced intelligent driving, offering all users a lifetime of free highway navigation assistance, ushering in an era of “universal advanced intelligent driving.”
The industry’s first lip movement recognition combined with an esports room
Along with intelligent driving, there’s also competition in the intelligent cabin space.
As early as October 2023, SIMO became the world’s first AI voice assistant to integrate Baidu's Wenxin Yiyan, applying a large AI model. SIMO boasts rapid recognition speeds and allows smooth interaction even in weak or no network conditions.
With the integration of a multimodal large model, SIMO 2.0 has added numerous capabilities, including fun explorations and role-playing. For instance, during the drive, users can ask SIMO about property prices in nearby neighborhoods, suggest a classic local eatery, or find a meeting spot for chatting with friends. SIMO 2.0 also features the industry's only lip movement recognition function, enhancing speech recognition accuracy in noisy environments to 90% by tracking users' lip movements.
Jiyue has also integrated lane-level navigation maps with intelligent driving SR information, creating a comprehensive reproduction of the physical environment. Elements such as greenery, waterways, detailed lane markings, crosswalks, traffic lights, and even the taillights of preceding vehicles are presented in a 1:1 ratio on the large screen, providing users with an immersive navigation experience.
The updated UI design is a significant highlight of this intelligent cabin upgrade, introducing a new futuristic theme called "Epoch Creation" to Jiyue’s fully immersive leisure cabin.
In terms of entertainment services and smart connectivity, the V2.0 version transforms the vehicle into a "mobile esports room," allowing users to play dozens of popular games like Black Myth: Wukong with the press of a button. Additionally, the cabin can connect seamlessly with Baidu's ecosystem, smartphones, wearable devices, and switch screens.
In a post-conference interview, when asked how Jiyue coped with its darkest moments over the past year, Xia Yiping became emotional several times.
As a premium intelligent automotive robotics brand, Jiyue has always been at the forefront of the smart wave, yet it has also faced criticism for its overly aggressive product design.
In this evolution, Jiyue focused its efforts on intelligent driving and the intelligent cabin. Xia Yiping believes that 2024 will be a pivotal year for the explosion of intelligent driving and AI technologies, where the entire intelligent driving industry will enter the era of pure vision and end-to-end large models, creating a watershed moment in competition—"If you can’t keep up, you’ll fall behind.”
Jiyue is therefore prepared for the upcoming boom alongside pure visual intelligent driving solutions.
