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
typed-decision-bench
4nt0ineB
Compares Jev, OpenJev, Laya, and other models on zero-shot classification tasks in English and French
crisis-triage
mazenDDr
Fine-tunes and evaluates Laya for multilingual disaster-message triage with uncertainty routing
layar
nullean
A .NET port of Laya provides ONNX and TorchSharp inference backends and CLI tools
LAYA-RLCD
shyamsridhar123
Teaches RLCD and trains Laya pilots through a tactical game, lessons, and benchmarks
laya-test
JVMoreiraD
Evaluates the Laya model's decision-making capabilities
laya-serve
c4bbage
Serves Laya inference with Go, dynamic batching, and TensorRT or CUDA execution
laya-mlx
Gates-456
Runs open-weight Laya typed decisions locally on Apple Silicon using MLX
Laya-Experiments
alperunlu07
Collects reproducible experiments using Laya, including a Pong controller comparison
laya-cpp
samiul000
Implements native C++ Laya inference with ONNX Runtime, benchmarking, and INT8 quantization
laya-doom
dylanbstorey
Runs Laya in a real-time control loop to play Doom on Apple Silicon
laya-dino
that-daniel
Controls Chrome's dinosaur game with Laya through a real-browser decision loop
poc-laya
MayukhCars24
Measures Laya classification latency with a FastAPI backend, web console, and observability tools
laya-ts
mikeboe
Serves the Laya model over an API and benchmarks it against the hosted Jev API
migration-laya
ViniCarvalhoDados
Evaluates Laya for triaging legacy SQL queries before database migration
Laya-Navigator
Vann4799
Adapts Laya decisions to workflow-state classification and next-action prediction
jev-laya-explore
NachiketKandari
Evaluates the local Laya decision model alongside TypeSafe's hosted Jev
laya-local-lab
torresnicolas0
Reproduces local evaluations of three Laya variants across support, guardrails, RAG, and model selection
laya-mlx-benchmarks
jayluxferro
Benchmarks Laya typed-decision models running through the MLX inference port
laya-medical-finetune
Priyanshu-5257
Provides RLCD fine-tuning recipes and evaluations for Laya on medical decision tasks
laya-multilingual-dml
minicom365
Runs and benchmarks Laya Multilingual on AMD hardware using DirectML
Jev-vs-Laya
aarush-dhingra
Compares Jev and Laya as chess decision-makers in a local browser arena
laya-fit-check
JhouCode
A Python kit reproduces Laya's published benchmark and compares it with free baselines on custom data
jev-v-laya
rnunley
A held-out MMLU-Pro benchmark compares answer routing with local Laya and hosted Jev
laya-mlx-windows-cuda
oboroge0
Documents running the MLX Laya model on Windows with an NVIDIA GPU
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