On the afternoon of February 11, Flytech officially launched the "Spark X2" large model trained on fully domestic computing power. This upgrade marks a new stage in the algorithm and engineering innovation of China's domestic computing power models. It not only matches the world's top level in general capabilities but also accurately addresses the key demands for industry application.

General Capabilities: Fully Match the Leading Models

The Spark X2 has achieved a qualitative leap in core capabilities. In dimensions such as mathematics, logical reasoning, language understanding, and intelligent agents (Agents), its performance is comparable to the best global models. In addition, Spark X2 continues to strengthen multilingual processing, supporting more than 130 languages, and maintaining an industry-leading position in the comprehensive performance of languages in key trade regions such as Latin America and ASEAN.

Industry Focus: Evolving from "General" to "Practical"

Based on algorithm breakthroughs and high-quality industry datasets, Spark X2 enhances professional thinking chains through reinforcement learning, significantly improving the accuracy of industry applications:

  • Specialized Fields: Focuses on optimizing education, healthcare, automotive, and agent scenarios.

  • Core Advantages: Provides more professional and high-experience technical solutions, ensuring reliable support in industry applications.

Experience Now: All Platforms Launching Simultaneously

Currently, the latest capabilities of Spark X2 are updated on the Spark X2 Official Website and the App. At the same time, the new API for developers has also been launched on the Flytech Open Platform, helping enterprises rapidly iterate their applications.