Announcement
llama.cpp adds decision-model support through `/v1/systemone`
ggml-org says the endpoint scores supplied options and returns probabilities in a single forward pass instead of generating text.
TLDR
ggml-org says llama.cpp’s server now supports decision models through /v1/systemone. The endpoint takes a state and questions, then scores options for tasks such as routing requests or checking an agent’s work. Its October 2 article lists five supported models and says OpenJev could also read images at the time.
Combined views
80.8K
2 Sources, first seen ago