Tencent Integrates Kimi-K3.1-Pro Model into WorkBuddy Platform

Tencent has integrated the Kimi-K3.1-Pro language model into its WorkBuddy suite. The implementation enables a 1-million-token context window, aimed at enhancing performance in long-sequence tasks such as complex code generation, 3D modeling, and financial research.

Development of the Kimi-K3.1-Pro series has involved a significant increase in reinforcement learning compute power. Benchmarking data indicates the model performs at a level comparable to other industry-standard proprietary models in areas like CAD rendering and automated financial workflows. In specific testing environments, such as OfficeQA Pro v2, the model has demonstrated competitive accuracy.

Access to the new integration is currently prioritized for enterprise-level clients and individual subscribers due to resource constraints. Tencent has also deployed a high-speed variant of the model, which claims to increase output throughput by two to three times compared to the standard iteration.