In-Short
- Baidu introduces ERNIE 4.5 and ERNIE X1 AI models, offering advanced multimodal and reasoning capabilities.
- ERNIE 4.5 excels in understanding diverse content types and outperforms GPT-4.5 at a fraction of the cost.
- ERNIE X1 specializes in deep-thinking reasoning and tool use, enhancing Chinese Q&A, literary creation, and complex calculations.
- Both models are available for individuals through ERNIE Bot and for enterprise via Baidu AI Cloud’s Qianfan platform.
Summary of Baidu’s New AI Models
Baidu has made a significant leap in the AI industry by launching two innovative foundation AI models, ERNIE 4.5 and ERNIE X1. These models are designed to enhance the capabilities of multimodal and reasoning AI systems, and Baidu is making them available for free to individuals through ERNIE Bot. The company’s goal is to integrate these models into its product ecosystem, including Baidu Search and the Wenxiaoyan app, to improve user experiences.
ERNIE 4.5: A Multimodal Foundation Model
ERNIE 4.5 stands out as a new generation native multimodal foundation model that offers collaborative optimization across various modalities such as text, images, audio, and video. It has the unique ability to understand complex internet content, including memes and satirical cartoons. Baidu boasts that ERNIE 4.5 surpasses GPT-4.5 in several benchmarks while being significantly more cost-effective, priced at just 1% of GPT-4.5’s cost.
The model’s success is attributed to cutting-edge technologies like dynamic attention masking, multimodal mixture-of-experts, and knowledge-centric training data construction. These advancements contribute to its enhanced language understanding, generation, reasoning, and memory capabilities.
ERNIE X1: Deep-Thinking Reasoning Model
ERNIE X1, on the other hand, is Baidu’s first multimodal deep-thinking reasoning model capable of tool use. It excels in Chinese knowledge Q&A, literary creation, and complex calculations. The model’s tool use includes advanced search, document Q&A, image understanding, AI image generation, and webpage reading. ERNIE X1’s capabilities are supported by technologies such as progressive reinforcement learning and a unified multi-faceted reward system.
For enterprise users and developers, ERNIE 4.5 is accessible through APIs on Baidu AI Cloud’s Qianfan platform, with competitive pricing structures. ERNIE X1 will soon follow suit. Baidu is optimistic about the future of AI and is committed to investing in AI, data centers, and cloud infrastructure to further develop its AI capabilities and next-generation models.
For more detailed information, please visit the original source.
