• Baidu has launched two new AI models, including ERNIE X1, which it claims rivals DeepSeek's R1 at half the cost.
  • ERNIE 4.5 offers enhanced multimodal understanding, with improved language, logic, and memory skills.

Baidu has announced the release of two new artificial intelligence models, including a reasoning-focused model it claims matches the performance of DeepSeek's R1 model at half the cost. The launch comes as Baidu seeks to strengthen its position in the highly competitive AI industry.

The first model, ERNIE X1, is designed with enhanced reasoning capabilities and the ability to use tools autonomously. 

"ERNIE X1 delivers performance on par with DeepSeek R1 at only half the price," Baidu said in a statement. 

The company highlighted the model's strengths in understanding, planning, reflection, and evolution.

The second model, ERNIE 4.5, is described as a foundation model with improved multimodal understanding. Baidu stated that ERNIE 4.5 has more advanced language skills and enhanced abilities in understanding, generation, logic, and memory. The model is also said to have a "high EQ," making it more adept at interpreting internet memes and satire.

Baidu was one of the first Chinese tech companies to launch a ChatGPT-style chatbot, but it has faced challenges in gaining traction for its Ernie large language model. Despite claiming performance comparable to OpenAI's GPT-4, Baidu continues to navigate stiff competition from domestic and international AI developers.

The release comes amid increasing pressure from DeepSeek, a Chinese AI startup that has introduced models it claims are comparable to or better than leading U.S. models at a lower cost. This has intensified competition in the AI sector, with companies racing to develop more capable and cost-effective models.

Multimodal AI systems like ERNIE 4.5 are capable of processing and integrating various types of data, including text, video, images, and audio, and can convert content across these formats. Baidu’s latest models reflect the growing emphasis on reasoning and multimodal capabilities in the AI industry.

Edited by Harshajit Sarmah