Text Classification • 0.1B • Updated • 297 • 4
\n#
\n New in llama.cpp: Decision Models/v1/systemone endpoint. You send a state (text, JSON, a screenshot) and typed questions. The model returns a probability for each option in a single forward pass.\nThe API follows the System One format introduced with TypeSafe's Jev model, so existing clients only need a new base URL. Implementation details are in PR #29818.
What is a decision model? A model that answers by scoring the options you give it, instead of generating text. A chat model spends one forward pass per output token, and its output still has to be parsed. A decision model reads the input once, and its answer is always one of your options, with a probability attached. Typical uses are routing a request, moderating content, checking that an agent's step worked, or choosing its next action.
\n##
\n Supported models| Model | Size | Based on | Languages | Images | License | Speed* |
|---|---|---|---|---|---|---|
| Julia-1 | 144M | mmBERT-small | 50+ | no | Apache 2.0 | 3 ms |
| Laya | 421M | ModernBERT-large | English | no | Apache 2.0 | 5 ms |
| Kev-4B | 4B | Qwen3.5-4B-Base | English | no | Apache 2.0 | 12 ms |
| lev | 4B | Qwen3.5-4B | English | no | Apache 2.0 | 36 ms |
| OpenJev | 27B | Qwen3.8-27B | en, de, fr, hi, zh, ja | yes | CC BY-NC 4.0 | 43 ms |
<sub>*Median time to answer one question, on one NVIDIA RTX PRO 6000.</sub>
\nFind these models in the Decision models collection, with more coming. The community Decision Index shows how they compare.
\n##
\n Quick startGet the latest llama.cpp from llama.app (or run llama update), then start a model:
llama serve -hf ggml-org/Kev-4B-GGUFA request contains a state and one or more questions. There are three question types:
\n| Type | You send | You get |
|---|---|---|
choice | options, with optional descriptions | the top option, plus a probability per option |
score | 2 to 10 levels, lowest first | the expected level (can fall between two) |
noul | a yes/no question | the probability of yes |
Send a request with your state and questions:
\ncurl http://localhost:8080/v1/systemone \\\n -H "Content-Type: application/json" \\\n -d '{\n "state": "Customer message: I was charged twice for my order last week and nobody has replied.",\n "questions": {\n "route": {\n "type": "choice",\n "instructions": "Which team should handle this?",\n "criteria": {\n "billing": "payments, charges, refunds, invoices",\n "shipping": "delivery, tracking, lost or late parcels",\n "technical": "bugs, errors, login problems"\n }\n },\n "angry": {\n "type": "noul",\n "instructions": "Is the customer angry?"\n },\n "urgency": {\n "type": "score",\n "instructions": "How urgent is this?",\n "criteria": ["can wait", "this week", "today", "right now"]\n }\n }\n }'Response (values rounded):
\n{\n "model": "ggml-org/Kev-4B-GGUF",\n "answers": {\n "route": {\n "type": "choice",\n "choice": "billing",\n "probabilities": {"billing": 0.9049, "shipping": 0.0275, "technical": 0.0676},\n "confidence": 0.8574\n },\n "angry": {\n "type": "noul",\n "noul": 0.8208\n },\n "urgency": {\n "type": "score",\n "score": 2.2821,\n "legend": {"0": "can wait", "1": "this week", "2": "today", "3": "right now"},\n "probabilities": {"0": 0.036, "1": 0.1937, "2": 0.2225, "3": 0.5478},\n "confidence": 0.2821\n }\n },\n "usage": {"input_tokens": 130, "output_tokens": 0}\n}The full reference is in the server docs.
\n##
\n ImagesSome models (OpenJev at the time of writing) can also read images, such as documents or screenshots. The vision projector downloads automatically:
\nllama serve -hf ggml-org/OpenJev-GGUFFor example, to classify an uploaded document:
\nimport base64\nimport requests\nwith open("document.png", "rb") as f:\n image = "data:image/png;base64," + base64.b64encode(f.read()).decode()\nresponse = requests.post("http://localhost:8080/v1/systemone", json={\n "state": "A file uploaded by a customer.",\n "images": [image],\n "questions": {\n "kind": {\n "type": "choice",\n "instructions": "What kind of document is this?",\n "criteria": {"invoice": None, "receipt": None, "contract": None, "other": None},\n },\n },\n})\nprint(response.json()["answers"]["kind"]["choice"]) # invoiceThe state can also be a list of chat messages. Any image_url part (data URL) is read as an image, same as chat completions.
##
\n Several models, one serverIn router mode, models load on demand and you pick one per request:
\nllama serve\ncurl http://localhost:8080/v1/systemone \\\n -H "Content-Type: application/json" \\\n -d '{"model": "ggml-org/Julia-1-GGUF:Q8_0", "state": "...", "questions": {...}}'/v1/models lists the ids. With a single model loaded, the model field is ignored.
##
\n Tipsshipping with bare labels, and tobilling (0.99) once each option had a description.llama serve -hf ggml-org/Kev-4B-GGUF:Q8_0 .##
\n What's nextNew open decision models come out every week, and we'll keep adding the best ones. Cloudflare's Clef is next. Is there one you particularly want? Tell us in the comments.
","headings":[]}}