IT Home reports llama.cpp and ggml reach version 1.0.0 with clearer API and ABI compatibility rules

IT Home reported that Georgi Gerganov announced on X that the open-source AI inference project llama.cpp and the underlying ggml engine have reached library version 1.0.0. The project says future major-version bumps will signal API or ABI breaking changes, helping developers judge upgrade risk. It also plans to expand CI with Hugging Face’s HF Buckets and HF Jobs, while broadening self-hosted test coverage across NVIDIA, AMD, Qualcomm, Arm, IBM and Intel hardware.