Huawei Cloud Pangu Large Model 5.0 Officially Launched: Upgraded Multimodal Capabilities, Offering Parameters from Hundreds of Millions to Trillions

At the Huawei HDC 2024 Developer Conference held today, Huawei’s Executive Director and CEO of Huawei Cloud, Zhang Ping’an, announced the official launch of Pangu Large Model 5.0, which has been upgraded in three aspects: full series, multimodal, and strong thinking.

  • Full Series: Pangu Large Model 5.0 includes models of different parameter sizes to adapt to various business scenarios.
  • Multimodal: Pangu Large Model 5.0 can better and more accurately understand the physical world, including more modalities such as text, images, videos, radar, infrared, and remote sensing. In terms of generation, Pangu 5.0 can generate multimodal content that conforms to the laws of the physical world, allowing innovation to be effortless.
  • Strong Thinking: Complex logical reasoning is the key for large models to become industry assistants. Pangu Large Model 5.0 combines chain-of-thought technology with strategic search technology, greatly enhancing mathematical capabilities and complex task planning capabilities.

In terms of the full series, Pangu Large Model 5.0 includes parameters ranging from hundreds of millions to trillions, providing Pangu Natural Language Large Model, Multimodal Large Model, Vision Large Model, Prediction Large Model, Scientific Computing Large Model, etc.

  • Pangu E Series: 1 billion parameters, capable of supporting intelligent applications on mobile phones, PCs, and other end-side devices;
  • Pangu P Series: 10 billion parameters, suitable for low-latency, low-cost inference scenarios;
  • Pangu U Series: Available in 135 billion and 230 billion parameters, suitable for handling complex tasks and can serve as the foundation for enterprise-wide large models;
  • Pangu S Series: Trillion-level parameters, capable of handling cross-domain multi-tasks, helping enterprises to better apply AI technology across the board.

Pangu Large Model 5.0 brings three major innovative upgrades, with significant enhancements in multimodal interaction, such as the addition of understanding capabilities for 10K ultra-high-resolution images.

In terms of content generation, Pangu Large Model adopts the industry’s first STCG technology, focusing on scenarios in industries such as autonomous driving, industrial manufacturing, and construction, capable of generating multimodal content that better conforms to physical laws.

In addition, Huawei has collaborated with Leju (Shenzhen) Robotics Technology Co., Ltd. to develop the Pangu Embodied Intelligence Large Model. According to the introduction, Huawei’s Pangu Embodied Intelligence Large Model has the ability to have multiple forms with one brain, enabling billions of end-side devices to have autonomous intelligence.

During the conference, Huawei also demonstrated the operation of the Pangu Embodied Intelligence Large Model. Leju’s “Kuafu” robot, leveraging the capabilities of the Pangu Large Model, could smoothly complete interactive demonstrations such as identifying objects, answering questions, high-fiving, and handing over water, directly showcasing the cooperative achievements based on the Pangu Large Model.

The “Kuafu” humanoid robot can complete complex task planning of more than 10 steps and handle multi-scenario generalization and multi-task processing during task execution. At the same time, the Pangu Large Model can also generate training videos required by robots, allowing them to learn various complex scenarios more quickly.

Zhang stated that in addition to humanoid robots, the Pangu Embodied Intelligence Large Model can also empower various forms of industrial robots and service robots, enabling them to help humans engage in dangerous and arduous work. “As everyone expects, let AI robots help us with laundry, cooking, and sweeping, giving us more time to read, write poetry, and paint.”

It is reported that over the past year, Huawei Cloud’s Pangu Large Model has been implemented in more than 30 industries and over 400 scenarios, playing a significant role in areas such as government affairs, finance, manufacturing, pharmaceutical research and development, coal mining, steel, railways, autonomous driving, industrial design, architectural design, and meteorology.

With the support of Pangu Large Model 5.0, Huawei also introduced a new XiaoYi assistant, which is more intelligent, smoother, and more humanized. It can tailor your travel mode based on the weather outside and help you take precautions. Moreover, XiaoYi can continuously learn to better adapt to your usage habits.

In addition, the XiaoYi intelligent body is integrated with the navigation bar, omnipresent and ready to be summoned at any time. Simply “feed” text, images, or documents to XiaoYi, and it can handle text, recognize images, and analyze documents conveniently and efficiently.