Tencent’s newly released 2025 R&D Big Data Report paints a radically different picture of software development at one of the world’s largest technology firms. The company announced that artificial intelligence now contributes to half of all newly written code.
Its in-house AI assistant, CodeBuddy, is used by more than 90% of engineers across hundreds of development teams, all powered by Tencent’s HunYuan large language model.
The timing of the announcement, made on China’s 1024 Programmer Day, underscores the scale of AI integration.
With 76% of Tencent’s global workforce currently engaged in research and development, the company is effectively operating as one of the world’s largest R&D organizations.
Each month, it generates 325 million new lines of code, fulfills over 370,000 development requests, and completes 25.2 million software builds.
How AI reshaped Tencent’s development backbone
Tencent’s report identifies AI as the new engine driving productivity across all development stages. Average coding times have been reduced by 40%, while overall R&D efficiency has improved by more than 20% compared to previous years.
This boost extends across business units ranging from WeChat and QQ to Tencent Cloud. The firm’s internal data show that AI plays a hands-on role from conceptualization to deployment.
CodeBuddy not only assists engineers in writing entire code segments but also generates documentation, test cases, and integration scripts, turning iterative development into a near-autonomous process.
Did you know?
Tencent’s developers write over 325 million lines of code each month, equal to the length of 10,000 average open-source projects combined.
What makes CodeBuddy central to Tencent engineering
At the heart of this transformation is CodeBuddy, Tencent’s proprietary AI programming assistant built on HunYuan. It now contributes 50% of all new code, whether responding to engineers’ prompts or completing suggestions in real time.
Its integration within Tencent’s ecosystem has reached a point where one in every two lines of new code includes direct AI participation.
CodeBuddy’s impact has gone beyond producing code snippets. It assists in large-scale application development, repository maintenance, and code conversion between frameworks.
Tencent Cloud reported that 65% of its new code originated from CodeBuddy, contributing to a 31.5% reduction in defects per thousand lines.
How automation scales across platforms like WeDev
Tencent’s R&D infrastructure is anchored by WeDev, an integrated platform that links hundreds of tools across teams. WeDev’s unified environment handles over 80 million daily data exchanges, cutting manual operations by 5.3 million per month and improving automation rates 67% year-over-year.
The platform’s optimization reduced the average turnaround time for requirements by 12 hours and supported 16,000 daily tasks.
The report highlights how cloud-based systems like AnyDev have reduced environmental preparation time from a full day to 1 minute.
Similar efficiencies were achieved in Tencent Games, where AI-driven art tools achieved 95% automation accuracy, and in WeChat Pay, where delivery cycle times were shortened by 31%.
ALSO READ | Supermicro launches 6U MicroBlade with AMD EPYC 4005
Why AI review and quality control are key to reliability
In addition to writing new code, artificial intelligence now drives Tencent’s quality assurance systems. The report reveals that 94% of code reviews involve AI pre-screening, allowing automated “quality inspectors” to flag potential issues before human intervention.
These systems identified 28% of all detected code defects and increased effective issue detection by 44%, leading to measurable improvements in reliability.
CodeBuddy also integrates static scanning tools backed by HunYuan’s contextual comprehension engine, helping engineers find and fix 5.4 million bugs and vulnerabilities annually.
Across Tencent’s teams, bug resolution time dropped by an average of 8 hours, transforming daily debugging routines.
Where Tencent’s R&D momentum is heading next
Tencent’s AI implementation strategy has moved beyond mere experimentation. Its framework couples generative systems with data-driven platforms to reach full-cycle automation in development workflows.
The company’s R&D investments reflect this shift. In the second quarter of 2025 alone, Tencent spent RMB 20.25 billion on research and innovation, a 17% year-over-year increase.
Executives expect that AI will eventually handle the majority of repetitive software creation while human engineers focus on complex architectures, security, and product vision.
As HunYuan continues to evolve and cross-platform integrations deepen, Tencent aims to transform its engineering network into a self-improving ecosystem that merges human expertise with machine precision.
In a statement within the report, Tencent characterized AI as the 'new engine of industrial software evolution.'
If its current trajectory continues, AI may soon become Tencent’s default collaborator, shaping the next generation of global digital infrastructure.
Comments (0)
Please sign in to leave a comment