RK3588
A flagship edge SoC from Rockchip featuring an integrated NPU with a nominal compute throughput of roughly 6 TOPS (INT8). It supports multiple AI inference frameworks and is widely used in development boards, industrial computers, and edge server scenarios.
Timeline (last 90 days)
The 2B version of Qwen 3.5 (about 2 billion parameters) has been successfully deployed locally on an RK3588 development board; it can handle text Q&A and single-image Q&A, and can be used through a web interface.
After the model is kept resident, pure text Q&A begins outputting in about 0.38 seconds, and short answers complete in about 0.7 to 1.4 seconds.
Both vision encoder and language model inference use the RK3588's NPU.
All Facts (6)
瑞芯微RK3588采用8nm工艺，集成四核A76+四核A55的大小核架构，内置6TOPS算力的NPU，支持8K视频解码
90%Unverified瑞芯微RK3588和联发科天玑系列已将AI推理能力下沉至百元级模组
75%Unverified瑞芯微RK3588芯片具有6 TOPS算力
55%UnverifiedAfter the model is kept resident, pure text Q&A begins outputting in about 0.38 seconds, and short answers complete in about 0.7 to 1.4 seconds.
50%UnverifiedBoth vision encoder and language model inference use the RK3588's NPU.
50%UnverifiedThe 2B version of Qwen 3.5 (about 2 billion parameters) has been successfully deployed locally on an RK3588 development board; it can handle text Q&A and single-image Q&A, and can be used through a web interface.
50%