Qingcheng Jizhi Launches Full-Link Token Solution for AI Infrastructure

Qingcheng Jizhi has introduced a comprehensive AI infrastructure solution, including the Chitu inference engine, to optimize token production and reduce deployment costs for domestic AI models in China.

Key points

• The Chitu inference engine enables low-bit width quantization, significantly lowering hardware requirements for deploying large models like DeepSeek-V3.
• The AI Ping platform provides a centralized hub for evaluating and routing over 600 large model services to improve cost-efficiency and performance.
• Qingcheng Jizhi's software stack aims to support the national 'Artificial Intelligence+' initiative by enhancing the autonomy and efficiency of domestic AI infrastructure.
• The company has secured funding from multiple state-backed and venture capital institutions, including Zhongke Chuangxing and CICC Capital.

Why it matters

This development addresses critical bottlenecks in China's AI industry, such as high inference costs and reliance on foreign engines, by fostering a self-reliant and efficient domestic AI ecosystem.