THE COMPLETE LIST

System One models

A System One model is a decision model: it reads a state and answers typed questions — choose an option, score on a scale, or give the probability that a condition holds — in one forward pass, with calibrated confidence and no generated text.

This page lists every System One model on the registry — 17 from 16 makers — with what each decides, its size, context, licence and price, and the benchmark numbers its publisher reports. Updated September 26, 2026.

What they answer

ModelDecidesParametersContextAvailability and priceLicenceReported accuracyMedian latencyReleased
gliner2-5-decideFastino Labschoice, score, noul, classify, extract, route340M—Open weights + hosted APIHosted, price not published, or free to self-hostapache-2.060.2%Fastino fast-decisions suite (17 datasets)38.3 msSep 24, 2026
clmContrastive-LMchoice, score, noul, rank, classify, route8.0B—Open weightsFree (open weights)apache-2.0——Sep 23, 2026
tev1Together AIchoice, classify, route4.0B—Open weights + hosted API$0.042 / $0 per 1M, or free to self-hostother88.0%Together development set (reused, not held out)—Sep 23, 2026
jevk5Alibi Serikbaychoice, score, noul, classify, route4.0B—Open weightsFree (open weights)apache-2.078.4%JevBench public hard tier13.2 msSep 22, 2026
jev-omniakhilaaa3choice, noul, score, classify12B—Open weightsFree (open weights)apache-2.087.6%DecisionBench Medium (author's set)83 msSep 20, 2026
openthai-systemoneiApp Technologychoice, score, noul, classify, route800M—Open weightsFree (open weights)apache-2.074.3%Bespoke public subsets (13, macro average)40 msSep 20, 2026
laya-multilingual-mlxaac6fefchoice, score, noul322M1K tokensOpen weightsFree (open weights)apache-2.0——Sep 19, 2026
laya-mlxaac6fefchoice, score, noul421M512 tokensOpen weightsFree (open weights)apache-2.0——Sep 19, 2026
vonwfzyxchoice, score, noul, classify395M8K tokensOpen weightsFree (open weights)apache-2.063.9%JevBench public standard tier18 msSep 19, 2026
kevJared Palmerchoice, score, noul, classify, route4.0B8K tokensOpen weightsFree (open weights)apache-2.083.8%transfer-v4 (locked, out of domain)—Sep 19, 2026
cua-s1-formsCuachoice706K—Open weightsFree (open weights)mit100.0%196-decision evaluation on real forms—Sep 18, 2026
bespoke-nimble-9bBespoke Labschoice, noul, score, classify, route9.0B8K tokensOpen weights + hosted APIHosted, price not published, or free to self-hostapache-2.090.1%Bespoke held-out set (324 examples)106 msSep 18, 2026
layaConvai Innovationschoice, score, noul, classify, route421M512 tokensOpen weightsFree (open weights)apache-2.0—39.5 msSep 18, 2026
nanojevTianyuCodingschoice, noul, score600M—Open weightsFree (open weights)mit——Sep 17, 2026
system-one-qwen3-5-4b-scorerpngwnchoice, score, noul, classify4.0B—Open weightsFree (open weights)cc-by-nc-4.070.7%author-reported112 msSep 16, 2026
deciderMapikachoice, score, noul, classify, route2.0B—Open weightsFree (open weights)apache-2.080.2%Decider 67-task regression set3.2 msSep 16, 2026
jevTypeSafe AIchoice, score, noul, classify, route—32K tokensHosted API$0.042 / $0 per 1Mproprietary——Sep 15, 2026

Accuracy and latency are what each publisher reports, on its own suite and hardware. They are not a ranking: the suites differ. Parameters, context, licence and price are from each model’s manifest.

Use one

Pull any open-weight model from your terminal. Files are verified against their SHA-256 and cached.

pip install systemonemodels
systemone pull convai-innovations/laya

Hosted models are called through their provider’s API; each model page has a playground that builds the request.

Fine-tune one

Laya Studio fine-tunes Laya on your own labelled data, on your own machine, measures it against the base model on the same held-out split, and publishes it here in one step.

The fine-tuning guide · What LoRA does

Who makes them

Head to head

Community fine-tunes

Questions

What is a System One model?

A System One model is a decision model: it reads a state and answers typed questions — choose an option, score on a scale, or give the probability that a condition holds — in one forward pass, with calibrated confidence and no generated text.

How many System One models are there?

The registry lists 17 System One models from 16 makers: 16 with open weights and 4 served as a hosted API, counting models that are both in each. People have published 5 fine-tunes and experiments on top of them.

What was the first System One model?

jev, from TypeSafe AI, released on Sep 15, 2026.

Which System One models are open source?

Open weights under a licence that allows commercial use: laya-multilingual-mlx (apache-2.0), laya-mlx (apache-2.0), jev-omni (apache-2.0), nanojev (mit), jevk5 (apache-2.0), cua-s1-forms (mit), openthai-systemone (apache-2.0), von (apache-2.0), decider (apache-2.0), bespoke-nimble-9b (apache-2.0), clm (apache-2.0), laya (apache-2.0), kev (apache-2.0) and gliner2-5-decide (apache-2.0). Some other models publish weights under a non-commercial licence or none at all, so check each model's licence before shipping it.

Which System One model is the fastest?

By the median latency their publishers report, the fastest are decider at 3.2 ms, jevk5 at 13.2 ms and von at 18 ms. The figures come from different hardware and different request sizes, so measure on your own setup before choosing on speed.

How much does a System One model cost?

Open-weight models are free to download and run on your own hardware. Hosted ones are billed by their providers per million input tokens: tev1 at $0.042 and jev at $0.042.

Which System One model is the most accurate?

There is no fair ranking yet. Every publisher reports accuracy on a different suite — their own held-out set, a JevBench tier, a domain benchmark — so the numbers in the table are not comparable with each other. The table shows the suite next to each figure; the reliable test is your own labelled examples.

How do I use a System One model?

Install the CLI with pip install systemonemodels, then systemone pull namespace/name downloads a model’s files, each verified against its SHA-256. Hosted models are called through their provider’s API, which each model page links to.

Can I fine-tune a System One model?

Yes. Laya Studio fine-tunes Laya on your own data, on your own machine, and publishes the result here in one step. Any open-weight model can be fine-tuned with its own training code and published with systemone push.

Read on: the glossary · What is a System One model? · System One models vs LLMs · The landscape, in depth