TCL Huaxing X-Intelligence large model continues to consolidate its global leadership in the display field

Recently, in automated evaluations and industry expert evaluations for the semiconductor display field, TCL CSOT's independently developed X-Intelligence large model's vertical domain understanding and in-depth reasoning capabilities outperformed international mainstream general-purpose large models such as Gemini 3.1 Pro and GPT-5.5 high, and continues to maintain its No. 1 position in the global display field.

In terms of understanding the cross-domain professional knowledge of semiconductor display, X-Intelligence addresses core technical issues such as process principles, equipment mechanisms, and material properties, and can output solutions that are more professional and more in line with industry realities than mainstream general-purpose large models. Relying on vertical multi-modal image and text analysis capabilities, X-Intelligence also performs outstandingly in process defect detection and abnormal countermeasures tasks. The accuracy of identifying various process defects in panel manufacturing and the quality of abnormal handling solutions are also better than mainstream general-purpose large models.

At the same time, in complex scenarios with strong reasoning such as process optimization that require a deep integration of physical mechanism logic and data-driven algorithms, Xingzhi X-Intelligence fully demonstrates its solid professional accumulation and in-depth scene understanding capabilities in the semiconductor display industry.

The X-Intelligence large model is the first vertical multi-modal large model with strong reasoning capabilities in the global display industry. It was jointly developed by TCL Huaxing and TCL Industrial Research Institute. Unlike general large models that rely on massive Internet data for training, X-Intelligence focuses on professional scenarios such as semiconductor display R&D and manufacturing, and builds a professional knowledge system covering the entire industry chain around core areas such as process knowledge, equipment mechanisms, and materials R&D, forming a set of sustainable and iterative data production and model training methods.

It is worth noting that the Xingzhi X-Intelligence large model is based on the Qianwen large model. By exploring the technical route of efficient vertical training, with a parameter scale of about 27B, it has surpassed the capabilities of the trillion-parameter-level general-purpose large model in professional semiconductor display tasks. This once again confirms that in industrial application scenarios, the competition for large models is gradually shifting from "competition in parameter scale" to "competition in industry knowledge, data quality and application capabilities". High-quality professional data and vertical domain training systems are becoming key factors in determining model capabilities.

Previously, the leading advantages and industrial value of the X-Intelligence large model have been continuously verified. In 2025, the X-Intelligence large model has ranked first in the semiconductor display field in the global industrial large model evaluation. This model has been deeply integrated into the entire process of TCL Huaxing's R&D, manufacturing and operations, becoming an important engine for the company's R&D innovation and smart manufacturing.

In the production and manufacturing process, the ADC and ADR technology based on this model realizes the entire process of "detection-judgment-repair" intelligent closed loop, and the accuracy of defect judgment is increased to 95%. In the field of R&D innovation, large models directly drive product and material development, helping to improve problem analysis and material R&D efficiency by 20%-30%. In the field of operational services, model capabilities are further extended to multiple scenarios such as personalized employee training and remote intelligent diagnosis, continuing to release the value of artificial intelligence empowering industrial development.

Currently, a new round of artificial intelligence technology represented by large models is accelerating its penetration into the manufacturing industry. The core competitiveness of large industrial models does not lie in the pursuit of larger parameter scales, but in the ability to penetrate into the front lines of the industry, accumulate industry knowledge, form sustainable iterative data capabilities, and truly solve complex problems in enterprise R&D, manufacturing and operations.

TCL Huaxing will continue to delve into independent innovation of large models in the display field, continue to contribute to the intelligent upgrading of the global display industry with its hard-core self-research capabilities, and assist the intelligent and high-quality upgrading of the global semiconductor display industry.