ZeroHour
DashboardNewsCVEsExploitsIncidentsAIBriefAnalyticsEntitiesWatchlistIOCsAsk the feedSourcesFeeds & APIKiosk mode
AI model

Laya

3 mentions in 7 days · 3 in 30 days · 3 total · first seen · last

Search everything

Timeline

Laya the open source version of Jev

Convai Innovations released Laya, an open-source Apache 2.0 family of non-autoregressive decision models claiming 6-8x faster inference than TypeSafe AI's Jev.

Convai Innovations launched Laya, a family of non-autoregressive 'System 1' decision models that output calibrated probabilities over schemas instead of generating text. It ships three Apache 2.0 checkpoints on Hugging Face: laya (421M, ModernBERT-large), laya-multilingual (322M, mmBERT-base, 100+ languages), and laya-typed-decisions (421M, 0.766 accuracy on security alerts), running in 32.8 ms on a single GPU (7.2 ms batched). The release positions Laya as an open alternative to TypeSafe AI's Jev, a commercial decision API priced at $0.042 per million input tokens with ~150 ms responses. The author cites prior arXiv papers (2503.23303, 2510.01237) and a 51-language MASSIVE benchmark sweep showing English-only encoders fail on non-Latin scripts (Khmer: 0.000 accuracy at 0.952 confidence).

Hacker News · security · 16h agoModel release 4 sourcesHN 22↑ · 3 comments

[AINews] Here are 6 Clones of Jev in 2 days

Jev's non-generative decision model launch drew 36M views and six open clones in two days, positioning discriminative models as fast complements to LLMs.

Latent Space's AI News roundup for 9/17-9/18/2026 centers on Jev, a non-generative 'System 1' decision model whose launch video hit 36M views in two days and reached roughly 13% of teams. At least six open reproductions appeared, including Bespoke Nimble (a Qwen3.5-9B LoRA fine-tune improving a curated eval from 66% to 90% versus Jev's 93%), SemIf (4B/35B Qwen3.5 backbones), and Kev-0.5B built on Qwen2.5-0.5B. Practitioners highlighted calibrated-probability use cases in routing, escalation, and browser/computer-use workflows, with Braintrust reporting ~400x lower scoring cost versus prior setups. The roundup also covers Claude Code v2.1.277 adopting the AGENTS.md standard and research showing coding-agent harness design materially shapes benchmark results.

Latent Spaceupdated · 16h agofirst · 21h agoAI industry 4 sources1

convaiinnovations/laya — new model trending #30 on Hugging Face

Convaiinnovations releases Laya, an open-source 421M-parameter non-autoregressive decision model for calibrated email triage, phishing detection, and routing.

Convaiinnovations published a fine-tuned checkpoint of Laya, an open-source (Apache 2.0) non-autoregressive System 1 decision model trending #30 on Hugging Face. It combines a fully fine-tuned ModernBERT-large backbone (395M) with a from-scratch decision head for 421M total parameters, returning typed answers with calibrated probabilities instead of generated text. Training used RLCD with proper scoring rules and TD(lambda=1.0) trajectory modeling on fully human-annotated data. Reported results include 38.4 ms single-question latency, 83.8% in-task macro accuracy, and 73.2% accuracy on email triage and phishing detection.

Hugging Face trending models · 1d agoModel release

Appears with

product Laya 2ai model Jev 1ai model Jevlike 1ai model Kev-0.5B 1ai model laya-multilingual 1ai model laya-typed-decisions 1ai model Qwen2.5-0.5B 1ai model Qwen3.5-9B 1ai model SemIf 1organization Convai Innovations 1organization Convaiinnovations 1organization Latent Space 1organization TypeSafe AI 1product AGENTS.md 1product Claude Code 1product Cline 1
Entities are extracted by the model from each article. Watching an entity keeps it in this browser only (no account); the watchlist page and dashboard alerts use it.