ggml-org (llama.cpp) · Open-weight decision model · Last verified 2026-10-15

llama.cpp decision endpoint (/v1/systemone)

Verified

llama-server in llama.cpp now serves a local System One-style /v1/systemone endpoint for open decision models.

What it does

  • Local /v1/systemone endpoint in llama-server, announced Oct 2, 2026.
  • Supported models: Julia-1 (144M), Laya (421M), Kev-4B, lev (4B), OpenJev (27B, CC BY-NC 4.0, images), Clef (27B).

Pricing

  • Free and open source; you pay your own compute. Check each model’s license (OpenJev is CC BY-NC 4.0, non-commercial). [Verified 2026-10-15 on Hugging Face blog]

Limits

  • Depends on your hardware and the model; no hosted SLA. [Verified 2026-10-15 on Hugging Face blog]

Status

  • Published Oct 2, 2026 on the Hugging Face blog by ggml-org. [Verified 2026-10-15 on Hugging Face blog]

Safety & fit notes

  • Run decision models on-prem behind the same request shape as hosted APIs.
  • Confirm licenses before commercial use.

Sources

Related