66 comparisons · checked 20 September 2026

Compare decision models

Head to head on the things that can actually be checked: licence, architecture, size, and whether anyone has measured it. Not on accuracy, because for most of these nobody has published a number worth repeating.

How to use these

Start with licence. It rules models out of a product faster than any benchmark and it is the one dimension that is never ambiguous. Then architecture, because it decides whether the confidence number is something you can route on. Size last.

Open model against open model

Neither side of these has an independent benchmark. What they do have is a licence, an architecture and a size, and those are enough to narrow a shortlist.

laya vs Qwen-2.5-1B-RLCDSame class, different sizelaya vs cua-s1-formsSame class, different sizelaya vs laya-multilingualSame class, different sizelaya vs decider-2bSame class, different sizelaya vs laya-typed-decisionsSame class, different sizelaya vs open-jev-deberta-v3-largeSame class, different sizelaya vs LFM2.5-350M-RLCDLicence differslaya vs modernbert-ja-310m-jevLicence differslaya vs JEV-CPUSame class, different sizelaya vs laya-vision-smolvlm-256mLicence differsQwen-2.5-1B-RLCD vs cua-s1-formsSame class, different sizeQwen-2.5-1B-RLCD vs laya-multilingualSame class, different sizeQwen-2.5-1B-RLCD vs decider-2bSame class, different sizeQwen-2.5-1B-RLCD vs laya-typed-decisionsSame class, different sizeQwen-2.5-1B-RLCD vs open-jev-deberta-v3-largeEncoder vs decoderQwen-2.5-1B-RLCD vs LFM2.5-350M-RLCDLicence differsQwen-2.5-1B-RLCD vs modernbert-ja-310m-jevLicence differsQwen-2.5-1B-RLCD vs JEV-CPUSame class, different sizeQwen-2.5-1B-RLCD vs laya-vision-smolvlm-256mLicence differscua-s1-forms vs laya-multilingualSame class, different sizecua-s1-forms vs decider-2bSame class, different sizecua-s1-forms vs laya-typed-decisionsSame class, different sizecua-s1-forms vs open-jev-deberta-v3-largeSame class, different sizecua-s1-forms vs LFM2.5-350M-RLCDLicence differscua-s1-forms vs modernbert-ja-310m-jevLicence differscua-s1-forms vs JEV-CPUSame class, different sizecua-s1-forms vs laya-vision-smolvlm-256mLicence differslaya-multilingual vs decider-2bSame class, different sizelaya-multilingual vs laya-typed-decisionsSame class, different sizelaya-multilingual vs open-jev-deberta-v3-largeSame class, different sizelaya-multilingual vs LFM2.5-350M-RLCDLicence differslaya-multilingual vs modernbert-ja-310m-jevLicence differslaya-multilingual vs JEV-CPUSame class, different sizelaya-multilingual vs laya-vision-smolvlm-256mLicence differsdecider-2b vs laya-typed-decisionsSame class, different sizedecider-2b vs open-jev-deberta-v3-largeEncoder vs decoderdecider-2b vs LFM2.5-350M-RLCDLicence differsdecider-2b vs modernbert-ja-310m-jevLicence differsdecider-2b vs JEV-CPUSame class, different sizedecider-2b vs laya-vision-smolvlm-256mLicence differslaya-typed-decisions vs open-jev-deberta-v3-largeSame class, different sizelaya-typed-decisions vs LFM2.5-350M-RLCDLicence differslaya-typed-decisions vs modernbert-ja-310m-jevLicence differslaya-typed-decisions vs JEV-CPUSame class, different sizelaya-typed-decisions vs laya-vision-smolvlm-256mLicence differsopen-jev-deberta-v3-large vs LFM2.5-350M-RLCDLicence differsopen-jev-deberta-v3-large vs modernbert-ja-310m-jevLicence differsopen-jev-deberta-v3-large vs JEV-CPUEncoder vs decoderopen-jev-deberta-v3-large vs laya-vision-smolvlm-256mLicence differsLFM2.5-350M-RLCD vs modernbert-ja-310m-jevEncoder vs decoderLFM2.5-350M-RLCD vs JEV-CPULicence differsLFM2.5-350M-RLCD vs laya-vision-smolvlm-256mSame class, different sizemodernbert-ja-310m-jev vs JEV-CPULicence differsmodernbert-ja-310m-jev vs laya-vision-smolvlm-256mEncoder vs decoderJEV-CPU vs laya-vision-smolvlm-256mLicence differs

The models in these comparisons

The 12 most-watched, excluding quantised builds and adapters. Comparing a model to its own GGUF conversion is not a comparison.

Jev

TypeSafe's hosted decision model. Returns a typed choice with a full probability distribution instead of text. The only model in this directory we have benchmarked end to end.

Not disclosed · Proprietary

laya

The most-liked open-weight decision model of the post-Jev wave, a 421M ModernBERT-large encoder. The base release of a three-model family.

421M · Not disclosed · apache-2.0

Qwen-2.5-1B-RLCD

Qwen2.5-1.5B-Instruct fine-tuned to return typed decisions.

Decoder · apache-2.0

cua-s1-forms

A System One model narrowed to one job, filling forms. MIT licensed, and the clearest example in the directory of the interface spreading past the benchmark it was born on.

Not disclosed · mit

laya-multilingual

The only multilingual decision model we have found, a 322M encoder on mmBERT. Also the one our English-only test set cannot say anything useful about.

322M · Not disclosed · apache-2.0

decider-2b

The most-downloaded open-weight decision model in the post-Jev wave. A Qwen3.5-2B decoder fine-tuned to emit typed decisions, and the largest member of a three-model family.

1.9B · Decoder · apache-2.0

laya-typed-decisions

The Laya variant tuned on the typed-decisions benchmark, and the only open model whose card publishes head-to-head numbers against Jev. One of those numbers is the reason we started measuring.

421M · Not disclosed · apache-2.0

open-jev-deberta-v3-large

A DeBERTa-v3-large encoder positioned as an open reimplementation of Jev. The highest ratio of likes to downloads in the directory, which is usually what a credible claim looks like early.

434M · Encoder · apache-2.0

LFM2.5-350M-RLCD

A typed-decision model from notnotsamuel, 354M parameters. Licensed other, check before commercial use.

354M · Decoder · other

modernbert-ja-310m-jev

modernbert-ja-310m fine-tuned to return typed decisions, 315M parameters. Licensed cc-by-sa-4.0, check before commercial use.

315M · Encoder · cc-by-sa-4.0

JEV-CPU

Qwen3-0.6B fine-tuned to return typed decisions.

Decoder · mit

laya-vision-smolvlm-256m

SmolVLM-256M-Instruct fine-tuned to return typed decisions, 237M parameters. Licensed cc-by-nc-sa-4.0, check before commercial use.

237M · Decoder · cc-by-nc-sa-4.0

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.