Chinese technology company Baidu has recently unveiled two new AI models within the ERNIE series of Large Language Models (LLM). The models are named ERNIE 4.5 and ERNIE X1. Additionally, the company has made its proprietary chatbot, Ernie Bot, free to use for all users.
“We’ve just unveiled ERNIE 4.5 & X1! As a deep-thinking reasoning model with multimodal capabilities, ERNIE X1 delivers performance on par with DeepSeek R1 at only half the price.”, the company posted on X (formerly Twitter). Baidu also released a short demo trailer showcasing some of ERNIE 4.5’s capabilities.
Earlier this year, Chinese AI start-up DeepSeek released the DeepSeek R1 model which generated widespread discussions in the industry. DeepSeek’s most notable aspect is its ability to match the performance of popular models such as ChatGPT and Gemini while costing a fraction of the price. ERNIE X1 is a reasoning model that can reportedly deliver similar results at even cheaper prices. This has already garnered much attention in the AI community.
See Related: China Sparks Huge Uproar In The AI Industry With Its Cutting-Edge And Cost-Effective AI Model
ERNIE 4.5, on the other hand, is a multimodal foundation model. While the X1 model is geared towards complex reasoning and logic-based tasks, the 4.5 is more akin to typical chatbot AI. According to Baidu, “ERNIE 4.5 has more advanced language ability, and its understanding, generation, logic, and memory abilities are comprehensively improved.”
Currently, both X1 and 4.5 models are available for free to all ERNIE bot users. ERNIE 4.5 input and output prices have been set at $0.55 per 1M tokens and $2.2 per 1M tokens, respectively. ERNIE X1 will cost $0.28 per 1M tokens for input and $1.1 per 1M tokens for output.
Baidu is a Chinese multinational tech company based in Beijing, China. The company specializes in internet services and AI tools. Baidu has followed several other Chinese AI companies in releasing an AI model that can compete with its American counterparts such as OpenAI, Google, Anthropic and others.
