gparitosh007/Open-Decision-Model-150M-ONNX

Open Decision Model 150M (ONNX fp16): calibrated decisions without PyTorch

The fp16 ONNX export of gparitosh007/Open-Decision-Model-150M, a small open decision model of the Jev / Laya type. Use it where PyTorch is too large, for example in a serverless function on Vercel.

Give the model a state and a list of options. It returns a calibrated probability for each option. It does not generate text. Use it for LLM routing, tool-risk checks, yes/no checks and other multiple-choice decisions.

Size	299 MB (fp16)
Runtime	ONNX Runtime; no PyTorch
Source model	Open-Decision-Model-150M, tag v2
Agreement with PyTorch	Max probability difference 0.000277 on 50 records
Licence	Apache 2.0 (commercial use is permitted)

Use it now: free public MCP endpoint

No download and no API key. A public, stateless instance runs on Vercel with this ONNX model: https://decision-model.vercel.app/mcp.

Claude Code:

claude mcp add --transport http decision https://decision-model.vercel.app/mcp

Other MCP clients (Claude Desktop, Cursor and others): add a remote HTTP MCP server with the same URL.

Tool	What it does	Required inputs	Returns
choice	Picks between options	state, instructions, criteria (one description per option)	A probability per option and a confidence
score	Rates on an ordered scale (for example trivial to very hard)	state, instructions, levels	A probability per level, the expected level and a legend
noul	Answers a yes/no question	state, instructions, optional yes/no definitions in criteria	p_yes
decide	Runs several questions against one state (Jev / Laya format)	state, questions (each with id, type, instructions, optional criteria)	One result per question
build_preferences	Asks a preference interview for one domain	domain, optional answers	A preference file for the domain
feedback	Records what you did with a decision	decision_id, action (accept, reject or other), optional chosen	A log of the outcome

All decision tools also accept domain (a saved preference file) or inline preferences, and question_id. The usual flow is: build_preferences, then choice, score or noul, then feedback.

The public instance is stateless: it saves no files and logs no request bodies. On it, send preferences inline in each call. To keep preference files and feedback, run your own server (see "How to use").

HTTP, without MCP:

curl -s https://decision-model.vercel.app/v1/decide -H 'Content-Type: application/json' -d '{
  "state": "Refactor the auth module and add tests. The repo has 40 files.",
  "questions": [{"id": "model", "type": "choice",
    "instructions": "Which model must handle this request?",
    "criteria": {"haiku": "Short and simple tasks", "sonnet": "Normal coding tasks", "opus": "Long or difficult tasks"}}]}'

The first call after an idle period takes about 50 s, because the function downloads the model. Later calls take less than 1 s.

Files

File	Content
model_fp16.onnx	Encoder and span head in fp16 (299 MB)
tokenizer.json	The tokenizer
decision_config.json	The input format and the calibration temperature
VERIFY.json	The check against the PyTorch model: max probability difference 0.000277 on 50 records

How to use

The input format, the option order and the output format are the same as in the PyTorch model. Read the model card of Open-Decision-Model-150M for the input layout, the training data, the evaluation and the limitations.

For AI agents

• Model id: gparitosh007/Open-Decision-Model-150M-ONNX. Model file: model_fp16.onnx.
• Use this repository for CPU or serverless inference. Use the PyTorch repository for fine-tuning.

Former name: decision-model-onnx. The old URL redirects here.