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
ai-decision-lab
uibuckets
Provides a local playground and benchmark harness for Laya, with optional Jev comparisons
IOCArena
hc-nolan
Compares Jev, Von, and Laya decision models on VirusTotal data
jev-zen
loongWoong
Reconstructs and validates Laya inference locally using NumPy and a browser-based interface
jev-eval-ja
unirt
Benchmarks Jev and Laya on Japanese business-style decision tasks
laya
harisathees
Tests Laya predictions, multilingual behavior, CPU performance, and interactive runs through a local web console
product_search_bench
Wanke15
Compares BM25, Jev, Qwen, and Laya for product search and reranking
laya-t-rex-runner
10086ggqq
A browser-based T-Rex runner uses typed-decision models and distilled agents to choose actions
laya-ROCm
don-milsey-miller
Provides an AMD ROCm runtime and benchmarks for running the Laya decision model
laya-vk
sulistta
Ports the Laya decision runtime to a Vulkan/IREE encoder backend for AMD and Qualcomm GPUs
laya-tictactoe
Kasa-Harendra
A terminal Tic-Tac-Toe demo runs the Laya model locally through Core ML on Apple Silicon
laya-onnx
Geoking2104
Runs Laya decision models with ONNX Runtime and provides export, inference, and benchmark tools
laya-agent
adhishthite
Benchmarks ConvAI Laya against TypeSafe Jev with and without live web grounding
laya-example
lim6112j
Demonstrates typed-question routing with Laya checkpoints and startup optimizations
laya-dino
JakkNaj
A fine-tuned Laya model plays Chrome Dino using a Python backend and TypeScript frontend
laya.cpp
kyr0
Implements native C++ inference and a JEV-compatible server for Laya typed decisions
laya-minesweeper
Animal2404
Demonstrates and evaluates Laya’s mine-risk judgments against a Minesweeper constraint solver
laya-snapdragon
piffie
Runs Laya typed decisions on Snapdragon X NPUs through ONNX Runtime and Qualcomm QNN
laya-tetris
kuchris
Runs Laya as a Tetris placement decision-maker with a heuristic safety guard
Laya_Playground
jonas050210
Runs local Laya decision demos, games, and a raw playground through a browser interface
laya-burmese
aungthuhein2005
Studies Laya zero-shot transfer, calibration, and fine-tuning for Burmese topic classification
laya-todo
firede
Classifies to-do items locally with Laya and compares its predictions with an optional Kev backend
laya-codex-bench
pbrehmer-ai
Benchmarks Laya-assisted Codex workflows for call reduction, latency, and decision quality
laya-gfx1030
Thotheris
Runs and benchmarks the Laya decision model on Windows with an AMD Radeon RX 6900 XT
laya-hexagon-npu
EricYu123456
Deploys Laya inference on Qualcomm Hexagon NPU using ONNX Runtime
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