Baidu to Launch Advanced AI Model Ernie 5.0 Amid Rising Competition
Chinese tech giant Baidu is set to launch its next-generation AI model, Ernie 5.0, in the second half of 2025.
This advanced model will feature multimodal capabilities, enabling it to process and convert information across text, video, images, and audio formats.
Key Points
- The launch of Ernie 5 occurs amid escalating competition in China’s AI sector, particularly startups like DeepSeek, which have developed cost-effective models rival established platforms such as OpenAI’s GPT.
- As an early entrant in AI with its Ernie model,Baidu has faced challenges in significant user engagement compared to competitors like ByteDance’s Bao and DeepSeek.
- Baidu announced that it would offer its AI chatbot, Ernie Bot, free of charge starting 1 april, 2025. This decision is attributed to advancements in technology and reduced operational costs.
- Baidu AI chatbot service will be accessible at no cost on both desktop and platforms.
- Additionally, Baidu plans to introduce an search function featuring expert-level responses and tool integration.
Background
Baidu’s Ernie (Enhanced through Knowledge Integration) is a pre-trained AI language model launched in 2019.
Initially developed for natural language understanding tasks, Ernie has evolved into multimodal AI capable of text, images, audio, and video.
The model leverages Baidu’s deep learning expertise large-scale knowledge graphs for enhanced contextual understanding.
Ernie has been progressively upgraded, with Ernie 4.0 and the upcoming Ernie 5.0 designed to improve chatbot performance, content creation, industry-specific applications in competitive AI.
News Gist
Chinese tech giant Baidu is set to launch its next-generation AI model, Ernie 5.0, in late 2025, featuring multimodal capabilities across text, video images, and audio.
Amid rising competition startups like DeepSeek, Baidu will offer its, Ernie Bot, free of charge starting April 1, 2025, along with a new search function providing expert-level responses and tool integration.