Laya alternatives
Jev was released in September 2026, and within weeks several open models with the same typed-question interface appeared. Laya is a 322M to 421M encoder. Most alternatives are open LLMs, mostly Qwen, fine-tuned at 0.5B to 35B parameters to score options without generating text. They trade speed for accuracy.
468 projects
laya-android
Byun11
Fine-tunes Laya into a typed-decision policy for Android GUI action selection
goal-classification-bench
fuljorge
Benchmarks Laya and Strands Decider for customer-goal classification in Brazilian Portuguese
miniLaya
yesterdayfuture
Fine-tunes a BERT-based typed-decision model and compares its outputs with Jev and Laya
decision-model-benchmarks
di37
Compares four decision models, including Laya, across classification, reranking, routing, and tool selection
business-classifier-benchmark
EduardoMoraga
Benchmarks classical classifiers against Jev and Laya for employee attrition prediction
laya-formatting-fragility
pranaysuyash
Measures how formatting changes affect typed-decision model outputs
laya-session-guard-risk-ranked-metrics
hxrikp
Publishes metrics from a rejected Laya prompt-injection risk-ranking experiment
extra-laya-bench-ft-data
cjdd3b
Provides additional data for Laya benchmark fine-tuning
decision-benchmark-jev-laya-julia
morrenhale
Provides benchmark data comparing decision models Jev, Laya, and Julia across classification tasks
laya-onnx-bench
gdelatournelle
Provides a reproducible evaluation set for Laya's ONNX decision backend
laya-browser-suite-c
cklxx
Provides held-out real-site tasks for evaluating a browser agent
next-jev-laya-test
JonesLin
Provides multilingual evaluation data for testing Jev and Laya across text-classification tasks
laya-nli-conflict-eval
slow-stack
A frozen evaluation set tests Laya's NLI memory-conflict head across a multi-run protocol
arc-usdc-laya-bench
goktugoguz
Publishes training rows, audited labels and benchmark results for Laya on Arc USDC transfer questions
riidolaya-statehint-laya-reference-v0.1-development
JooYoon
Reports development measurements using a frozen pretrained Laya checkpoint
stellar-laya-bench
goktugoguz
Publishes benchmark results and held-out evaluations for Laya on Stellar operations
laya-nli-conflict-eval
Modusnsus
Provides a frozen evaluation set for Laya's NLI memory-conflict decision head
laya-multidomain-decisions
devinblaze
A 665-example benchmark dataset of structured decisions across four domains
laya-rs
codesoda
A planned Rust runtime for local Laya decisions with Metal acceleration and a Jev-compatible HTTP API
laya-tetris-finetuning
hama-jp
Fine-tunes Laya for real-time Tetris placement decisions and documents the experiments
Laya-vs-Jev
abcdmku
A side-by-side test bench compares the local Laya model with the hosted Jev API
laya-blackjack
adambloebaum
Trains and evaluates a Laya-based blackjack policy with a simulator and probability dashboard
laya-car-snake
KarimHFahmy
Uses Laya to choose safe moves in a car-and-trailer grid game
laya-warehouse-safety
senthilr-nv
Simulates warehouse robot decisions using Laya, with a browser replay and deterministic collision-safety layer
The models
Sizes and licences are from each model's Hugging Face card.
| model | maker | sizes | licence |
|---|---|---|---|
| Kev | Jared Palmer | 0.5B to 27B (benchmarked: 0.8B, 4B, 9B) | Apache 2.0 |
| Decider | Mapika | 0.8B to 35B, plus a 2B vision model | Apache 2.0 |
| Clef, Clef-flash | Cloudflare | 27B, 9B | Apache 2.0 |
| GLiNER2.5-Decide | Fastino | 340M English, 287M multilingual, 1B | Apache 2.0 |
| Julia-1 | Supersonic Labs | 144M | Apache 2.0 |
| Jebadiah | Frontier Infra | 4B, 9B, 27B | Apache 2.0 |
| pplx-decider | Perplexity | ~26B ("27b" in the name) | Apache 2.0 |
- Kev runs a frozen Qwen3.5 base prefill-only, with a LoRA adapter and a pointer head over the options. The card says states of up to 65,536 tokens are served and 8,192 are validated.
- Mapika Decider calls itself an open reproduction of the System One model class. Its training code and data registry are on GitHub.
- Clef and Clef-flash read text, JSON, images and video, per their cards. Cloudflare's announcement says they are Jev-API compatible and also served on Workers AI.
- GLiNER2.5-Decide is a classification-only member of the GLiNER2.5 family: any label set at call time, single- or multi-label.
- Julia-1 is, like Laya, an encoder: it starts from mmBERT-small. See ModernBERT and mmBERT in Laya.
- Jebadiah ships a standalone server that answers
POST /v1/systemoneon one CUDA GPU or an Apple Silicon Mac. - pplx-decider v1.1 uses a separate decision readout instead of a normal language-model head, so its card asks you to use the included inference code.
The hosted reference is Jev from TypeSafe. Its weights are closed and it is billed per input token. The differences from Laya are in Laya vs Jev.
What independent benchmarks found
akitaonrails (October 10, 2026) used synthetic, AI-assisted cases in English and Brazilian Portuguese without independent human labels. Chance was about 31%. Local models ran on a Ryzen AI Max+ 395. Compact-set accuracy and median latency:
| model | accuracy | latency |
|---|---|---|
| Jev (hosted) | 97.6% | 255 ms |
| Clef (hosted) | 95.8% | 302 ms |
| Perplexity Decider 1.1 | 99.4% | 1,547 ms |
| Jebadiah 4B v2 | 90.5% | 205 ms |
| Kev 9B | 88.1% | 561 ms |
| Kev 4B | 87.5% | 246 ms |
| Mapika Decider | 87.5% | 186 ms |
| Clef-flash (local) | 87.5% | 618 ms |
| Kev 0.8B | 63.1% | 86 ms |
| GLiNER2.5-Decide (English only) | 52.4% | 208 ms |
| Laya (English) | 42.9% | 41 ms |
| Julia-1 | 31.5% | 12 ms (GPU) |
The author's conclusion: if Jev-like accuracy is a requirement, Laya, Julia-1 and the 0.6B Decision 3.0 Edge are not usable out of the box. He also notes that judging these small encoders by the accuracy column alone misreads them: they are built to be fine-tuned and run at the edge. Fine-tuning was not tested.
Arize (October 2026) ran a bracket of hallucination detection on RAGTruth, LLM-evaluator suites and agent routing. Balanced accuracy on RAGTruth was 85.5% for Jev, 85.2% for Clef 27B, 76.8% for Kev 9B and 59.8% for laya-multilingual. Laya was the fastest model at 0.04 s p50 on an M4 Max, and had the highest calibration error (0.330).
systemonemodels.ai lists 29 open Jev alternatives with the accuracy each publisher reports. Those figures come from different test sets, so they cannot be compared with each other.
How to choose
- Speed, CPU or on-device: the encoders. Laya and Julia-1 answer in tens of milliseconds, and Laya has the most ports, from Mac to the browser. Plan to fine-tune or to keep a fallback for hard cases.
- Accuracy without training: the 4B and larger models. They need a GPU or an Apple Silicon Mac with enough memory, and they are several times slower.
- No servers: Jev, Clef on Workers AI, or a hosted Laya API.
Whichever you pick, run a few hundred of your own labelled examples through two or three candidates. The rankings above come from small or synthetic sets, and their authors say so.
More ways to use Laya