OpenCV 5.0 Released With Rewritten DNN Engine, Built-In LLM & VLM Support

OpenCV 5.0 has been released as a major update to the popular open-source computer vision library. Key highlights include a fully rewritten deep neural network (DNN) engine, ONNX model coverage exceeding 80%, built-in support for large language models (LLMs) and vision language models (VLMs), a new hardware abstraction layer, and an improved 3D vision toolkit. Performance-optimized paths are available for Intel IPP (SSE/AVX), Arm KleidiCV, Qualcomm FastCV, and RISC-V Vector RVV. Native GPU support within the new DNN engine is planned for a future release. Benchmarks show OpenCV 5.0 performing competitively against Microsoft ONNX Runtime.