Baidu has launched the upgraded inference model ERNIE X1.1, claiming it rivals US models like OpenAI GPT-5 and Google Gemini 2.5 Pro in benchmark tests and surpasses the DeepSeek R1 update from May. At the WAVE SUMMIT, CTO Wang Haifeng revealed that X1.1 improves knowledge accuracy by 34.8%, instruction-following by 12.5%, and agent capabilities by 9.6% compared to ERNIE X1.