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.
472 projects
laya-qcs6490
nameissakthi25
Converts and benchmarks Laya on a Qualcomm QCS6490 NPU using a reproducible quantization recipe
laya-btc-decision-console
avtandilkasparov-debug
Measures Laya decision latency and compares paper-trading strategies in a local dashboard
laya-on-rk3588-jetson
suharvest
Converts and benchmarks Laya for deployment on Rockchip RK3588 and NVIDIA Jetson Orin
graph-loop-train
jtatman
Runs an adaptive training loop to improve fine-tuned Laya models
jev-laya-lab
luoshuai990529
Compares Jev and locally hosted Laya on decision accuracy, answer distributions, and task completion time
LayaETF
AM1975MA
Benchmarks Laya formulations for pairwise ETF decisions using a leakage-safe research harness
laya-reproduction
jimni3155
Reproduces selected Laya benchmark experiments, including typed-decision checks and option-name robustness
laya-retex
TojiiSamaa
Reports measured Laya experiments, fine-tuning, and applications in local search and tool routing
laya-cycle-dashboard
gchat137
Displays replayable simulation results from Laya decision-model game runs
laya-incident-triage
curtiscovington
Evaluates Laya against rule-based and trained baselines for incident triage and routing
jeff-laya-routing-study
tiw
Compares Jeff and Laya for sub-agent routing using a stacking model and confidence cascade
yt_laya_groq_comparison
BLUERAY94
Benchmarks Laya locally and optionally compares its checkpoints with Groq
decision-map-bench
move38studios
Benchmarks Jev, Clef, and Laya decision models on geography questions across a world grid
laya-check
e9tgldr
Benchmarks Laya-MN against JEV on Mongolian language understanding
laya-pii-bench
dfridkin
Benchmarks the open-weight Laya model as a local PII classifier on synthetic clinical-trial documents
Laya-sol
sushan5140
Prepares and validates an evaluation of GPT-6.1 Sol against Laya's typed-decision benchmark
laya-experiments
Mac-007
Runs experiments and empirical evaluations of Laya for text-based decision making
laya-trader
zakanysa
Tests Laya-driven trading strategies on an on-chain order book and evaluates their real costs
laya-benchmark
LokenSI
Benchmarks local decision models and Jev against Laya on English and Norwegian tasks
laya_mobile
SerAbin1
A Flutter app for testing Laya model performance on mobile
laya-fast-finetune
sathwikbs
Fine-tunes Laya with a packed training layout to reduce repeated state encoding
Jev-OpenSource-Alternative-Laya
Nagwa-Ahmed
Evaluates local Laya checkpoints on zero-shot text classification datasets and exports Excel reports
agent-failure-triage
PawnsApprentice
An experiment compares Laya with TF-IDF and an LLM judge for triaging failed agent runs
compare-IA-decision
laurentvv
A benchmark harness compares local System One decision models, including Laya, using measured results
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