Yeah, for example llama.cpp runs on Intel GPUs via Vulkan or SYCL. The latter is actively being maintained by Intel developers.

Obviously that is only one piece of software, but its a certainly a useful one if you are using one of the many LLMs it supports.