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
Decision-Models-Are-Not-Calculators
AnthusAI
A preregistered study compares Jev, Kev, and Laya on fair-die predictions
decision-model-eval
relevan-dev
An evaluation compares Laya, Jev, and Claude on search-index configuration decisions
doomgemma
dexmac221
Compares Laya with fine-tuned Gemma models in FreeDoom accuracy, calibration, latency, and gameplay tests
laya-email-classifier
mpegram3rd
Explores Laya configurations for classifying email
laya-e2-multitype
13974737898-lgtm
Studies Laya decision quality through reproducible benchmarks, comparisons, and fine-tuning experiments
maclaya
jannisleifeld
Serves local Laya models on Apple Silicon through a Jev-compatible API with a dashboard and playground
laya-experiments
RafaelJaime
Spanish notebooks evaluate Laya decision behavior, diagnostics, message routing, and product recommendation
laya-ludo
mnk786
A browser-based four-player Ludo game uses Laya and Jev to choose legal moves
laya-jev-homeautomation
CoryHawkless
A home-automation command router and benchmark harness exercises Laya
rag-laya-poc
Mr-PU
Compares standard RAG with a pipeline that uses Laya to rerank retrieved passages
laya-football-evaluation
steve-dev-55
Preregisters a multi-agent protocol for evaluating typed decision engines as real-time football forecasters
layev
CompleteDotTech
A research implementation of typed decisions, calibration, and reward training inspired by Kev and Laya
zh-decision-bench
CodyQin
Benchmarks calibration and robustness of System One decision models on Chinese-language tasks
system-one-benchmark
yanng981
Compares accuracy, calibration and latency across System One models on multilingual datasets
laya-android
edwardmonteiro
An Android space dogfight uses Laya’s typed decisions to control an enemy fighter
laya-mlx-mirror-227
Shijiuwei
Runs Laya inference locally on Apple Silicon using MLX, with a terminal Snake demo
clasificacion-tipada
tux22
Reproduces benchmarks comparing Jev, Laya, and local LLM classifiers on public datasets
laya-reasoning-demo
Prady0909
Demonstrates a recurrent reasoning extension of Laya with adaptive halting
Laya
mayataji3-star
Evaluates Laya on synthetic support tickets and preserves labels, predictions, code, and results
email_Jev_Laya
APRESIA
Benchmarks Laya, Jev and Qwen on email classification, including a Laya-to-Qwen cascade
layax
ohboyftw
Adds token-level option scoring and an abstention gate to Laya decision models
laya-swe
Mukul-svg
Uses Laya as a fast memory controller for autonomous coding agents and evaluates it on software tasks
laya-fintiment
shanaka95
Fine-tunes Laya for three-class financial sentiment and provides the training code, evaluations, and logs
laya-onnx
D-Ajay-Kumar
Converts Laya weights to ONNX and benchmarks the converted model
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