ONNX · 54 models available
Run a decision model with ONNX Runtime
An ONNX export loads without PyTorch. That matters for slim containers, for edge devices, and for running a decision inside a browser.
How to run it
Install the runtime
ONNX Runtime loads without PyTorch, which is usually the point. Container images get much smaller.
pip install onnxruntime huggingface_hubFetch the exported graph
ONNX repositories ship a .onnx file plus the tokeniser. Download both.
Consider the browser
ONNX Runtime Web and transformers.js run the same graph client side, which puts the decision on the user’s machine with no server call at all.
Models available in ONNX
Every entry in the directory that ships this format, read from its repository metadata on2 November 2026.
rlcd-modernbert-151m
The smallest decision model we track, at 151M parameters on a GLiClass/ModernBERT encoder. If it holds accuracy, it changes what this class of model costs to run.
bekko-system-one-v0-400m
An ONNX build of ettin-reranker-400m-v1, for running typed decisions on ONNX Runtime instead of a hosted API.
bekko-system-one-v0-17m
An ONNX build of ettin-reranker-17m-v1, for running typed decisions on ONNX Runtime instead of a hosted API.
laya-onnx
A ONNX build of laya, for running typed decisions on runtimes without PyTorch instead of a hosted API.
raya
laya fine-tuned to return typed decisions, 322M parameters.
laya-multilingual-ONNX
An ONNX build of laya-multilingual, for running typed decisions on ONNX Runtime instead of a hosted API.
laya-onnx
An ONNX build of laya, for running typed decisions on ONNX Runtime instead of a hosted API.
laya-onnx
An ONNX build of laya, for running typed decisions on ONNX Runtime instead of a hosted API.
laya-multilingual-onnx
An ONNX build of laya-multilingual, for running typed decisions on ONNX Runtime instead of a hosted API.
laya-onnx-fp16
A ONNX build of laya, for running typed decisions on runtimes without PyTorch instead of a hosted API.
wolf-defender-jev
mmBERT-small fine-tuned to return typed decisions, 151M parameters.
bekko-system-one-v0-68m
An ONNX build of ettin-reranker-68m-v1, for running typed decisions on ONNX Runtime instead of a hosted API.
laya-multilingual-onnx-int4
An ONNX build of laya, for running typed decisions on ONNX Runtime instead of a hosted API.
decider-12b-NPU-LPBQ-X-Elite
An ONNX build of decider-12b, for running typed decisions on ONNX Runtime instead of a hosted API.
laya-onnx
An ONNX build of laya, for running typed decisions on ONNX Runtime instead of a hosted API.
laya-onnx
An ONNX build of laya, for running typed decisions on ONNX Runtime instead of a hosted API.
laya-cybersec
laya fine-tuned to return typed decisions, 322M parameters.
laya-ko
laya fine-tuned to return typed decisions, 322M parameters.
laya-onnx-q8
An ONNX build of laya, for running typed decisions on ONNX Runtime instead of a hosted API.
laya-web-q8
An ONNX build of laya, for running typed decisions on ONNX Runtime instead of a hosted API.
laya-onnx
An ONNX build of laya, for running typed decisions on ONNX Runtime instead of a hosted API.
open-jev-laya-multilingual-onnx
open-jev-laya-multilingual-onnx, an open-weight model published to return typed decisions.
recmeets-laya-words
laya fine-tuned to return typed decisions.
laya-multilingual-onnx
An ONNX build of laya-multilingual, for running typed decisions on ONNX Runtime instead of a hosted API.
laya-typed-decisions-onnx
An ONNX build of laya-typed-decisions, for running typed decisions on ONNX Runtime instead of a hosted API.
cua-s1-forms-onnx
An ONNX build of cua-s1-forms, for running typed decisions on ONNX Runtime instead of a hosted API.
cut-laya-onnx
An ONNX build of laya, for running typed decisions on ONNX Runtime instead of a hosted API.
cyber-jev
nano-jev fine-tuned to return typed decisions, 23M parameters.
daylens-laya-onnx
An ONNX build of laya, for running typed decisions on ONNX Runtime instead of a hosted API.
jev-d-0.4b-onnx
laya-typed-decisions fine-tuned to return typed decisions.
laya-multilingual-onnx-int8
An ONNX build of laya-multilingual, for running typed decisions on ONNX Runtime instead of a hosted API.
laya-onnx-int8
An ONNX build of laya, for running typed decisions on ONNX Runtime instead of a hosted API.
laya-typed-decisions-web-q8
An ONNX build of laya-typed-decisions, for running typed decisions on ONNX Runtime instead of a hosted API.
layaForWeb
An ONNX build of laya, for running typed decisions on ONNX Runtime instead of a hosted API.
layaForWebTrained
An ONNX build of laya-typed-decisions, for running typed decisions on ONNX Runtime instead of a hosted API.
open-jev-deberta-v3-large-ONNX
An ONNX build of open-jev-deberta-v3-large, for running typed decisions on ONNX Runtime instead of a hosted API.
laya-onnx
An ONNX build of laya, for running typed decisions on ONNX Runtime instead of a hosted API.
laya-onnx
An ONNX build of laya, for running typed decisions on ONNX Runtime instead of a hosted API.
decider-0.8b-lora-onnx
An ONNX build of decider-0.8b, for running typed decisions on ONNX Runtime instead of a hosted API.
jev-gemma-4-E2B-it-choice-64
An ONNX build of gemma-4-E2B-it, for running typed decisions on ONNX Runtime instead of a hosted API.
JevK5-Lite-onnx
An ONNX build of JevK5-Lite, for running typed decisions on ONNX Runtime instead of a hosted API.
laya-bangla
laya-multilingual fine-tuned to return typed decisions, 322M parameters.
laya-bangla-lora
A LoRA adapter that turns laya-multilingual into a typed-decision model without shipping full weights.
laya-swedish-preference-gate
laya fine-tuned to return typed decisions, 322M parameters.
laya-typed-decisions-ONNX
An ONNX build of laya-typed-decisions, for running typed decisions on ONNX Runtime instead of a hosted API.
laya-ONNX
An ONNX build of laya, for running typed decisions on ONNX Runtime instead of a hosted API.
laya-typed-decisions-ONNX
An ONNX build of laya-typed-decisions, for running typed decisions on ONNX Runtime instead of a hosted API.
strands-decider-2B-hobson-v19-ONNX
An ONNX build of strands-decider-2B-hobson-v19, for running typed decisions on ONNX Runtime instead of a hosted API.
strands-decider-2b-web
An ONNX build of strands-decider-2B-hobson-v19, for running typed decisions on ONNX Runtime instead of a hosted API.
system-one-qwen3.5-4b-scorer-ONNX
An ONNX build of system-one-qwen3.5-4b-scorer, for running typed decisions on ONNX Runtime instead of a hosted API.
system-one-qwen3.5-4b-scorer-ONNX-etheroi
An ONNX build of Qwen3.5-4B-Base, for running typed decisions on ONNX Runtime instead of a hosted API.
laya-multilingual-onnx
An ONNX build of laya-multilingual, for running typed decisions on ONNX Runtime instead of a hosted API.
laya-onnx
An ONNX build of laya, for running typed decisions on ONNX Runtime instead of a hosted API.
verdict-laya
laya fine-tuned to return typed decisions, 421M parameters.
Also browsable as a facet of the directory: ONNX builds.
Other ways to run one
GGUF
GGUF is the format that made local inference ordinary. For decision models it matters more than usual, because the whole argument for the category is cost per call.
MLX
MLX uses unified memory on M-series chips. For a model in the few-hundred-megabyte range this is the fastest local option on a Mac, and it needs no GPU of its own.
Core ML
Core ML puts the model on the Neural Engine, so the decision happens on the phone or the laptop and nothing leaves it. The strongest privacy story in the category.
Compatible servers
16 open projects that reimplement the decision API, if you want the interface rather than just the weights.
Find out what your data actually supports
We turn your historical decisions into a frozen test set, then tell you which model, which threshold, and how much of it you can safely automate.