Company
Company behind Jev, a small model that outputs only structured responses like yes/no probabilities, used for guardrail and injection-detection tasks.
Sayora's receptionist test found Jev gave the LLM a wrong page on 3% of 168 questions, against 33% for its embeddings. The figure is Sayora's own, from four small-business sites, and transfers only to sites built from single-topic pages.
Reality
Subquadratic's SubQ Snap decision model accepts 384,000 tokens per request, 12 times Jev's 32K limit, according to the company. CEO Alex Whedon says the cap reflects preview compute capacity. The accuracy and latency figures are Subquadratic's own.
Reality
Workday gave TypeSafe's Jev every atomic component in its design system and lets the model pick what Sana shows, said chief AI officer Joel Hellermark. TypeSafe claims 70 to 500 milliseconds per answer; whether Workday sees that depends on how many screens reduce to closed questions.
Reality
Microsoft has put Decision-1, a model that scores fixed choices, in Foundry at $0.042 per million input tokens. Agent teams now have to test whether its scores, so far benchmarked only by Microsoft, are calibrated well enough to decide when a case goes to a human.
Perspective Coverage
7 publishersReality
Cloudflare's Clef-omni takes audio and video alongside text and images and returns schema-bound decisions in one API call. Runtimewire reported that the same release cut hosted Clef-flash's context window from 64,000 to 24,000 tokens, so longer prompts now go to Clef or to self-hosted weights.
Perspective Coverage
3 publishersReality
Cloudflare and AWS have followed TypeSafe's Jev with Clef and Strands Decider 2B, small models built to pick an agent's next tool or step. The pitch is a smaller token bill, though the practitioners dev.to cites say calibration and the cost of a wrong pick decide whether any saving survives.
Reality
OpenAI's Decisions API answers yes/no, pick-one and scale questions for $0.10 per million input tokens, with output free. Jev sells the same three question types for text at less than half that, so OpenAI's case rests on image input and compliance terms.
Perspective Coverage
4 publishersReality
Ollama and llama.cpp added local endpoints within a week for Jev decision models, 144M-to-27B checkpoints that return a label and a probability. Routing calls can move off hosted models once teams check each checkpoint's licence and test it on their own labels.
Reality
TypeSafe's Jev 1.13 matched no-CoT GPT-5.5 and Opus 4.7 on agent-monitoring AUROC at about 1/200th of Opus 4.7's cost, a LessWrong benchmark found. It answers in one step without writing text, so its calls leave no reasoning trace to audit.
Reality
LDraw Nova, Carlos Antelo's open-source tool, had Claude Opus 5.5 design a 2,175-piece LEGO garden by writing a Python program that emits the CAD file. Nova checks parts for collisions but not stability, so whether any of its designs would stand up in real bricks is still untested.
Reality
Ollama 0.35 adds a /v1/systemone endpoint returning a choice, yes/no or score from models run on the device. Its 9B Nimble model matched human moderation labels 70.3% of the time in its maker's test, so each team still sets its own review threshold.
Reality
Featherless open-sourced Simple Jev, a library that has open models pick fixed labels or yes/no answers, with hosting from $0.03 per million input tokens. CEO Eugene Cheah pitches it for classification work teams now pay frontier-model prices to run.
Reality
Jev's own gateway benchmark shows routing raised Opus 5 input tokens 61% on a Claude Code feature task, where the gateway can only hint at tools. Any saving depends on the task and on how many tools Claude Code sends the router each turn.
Reality
beans-picker cut a median macOS agent task from $0.388 to $0.060 in its author's benchmark by replacing full UI-tree reads with short candidate lists. Whether that carries over depends on how much of your own agent's bill goes to reading the window.
Reality
Mert Cobanov's open-source Ollaya serves local decision models behind a Jev-compatible API and reports 8 to 10 ms per request on an RTX 4090. Its comparisons with TypeSafe's hosted service are uncontrolled, so teams have to check accuracy on their own traffic before switching.
Reality
One dev.to author's Playwright gate has a fast model answer three typed questions per browser action, then lets plain code return allow, ask or block. The labelled evaluation is still unfinished, so the gate's unsafe-allow rate is unmeasured.
Reality
An outsider benchmarking laya's multilingual checkpoint on 300 Japanese emails found the first-listed ordinal level effectively unreachable, and the maintainer shipped the limitation in a README before any retrain.
Reality
A pre-registered run reads Gemma's label logits on one 24 GB GPU and scores it against the hosted API on the same 3,880 records, where it is level on yes/no questions and 4.5 points behind on multiple choice. One temperature fitted on 50 labels closes the calibration gap.
Reality
The model writing each page's recommendations only ever saw the candidates the code handed it, and when similarity matching found nothing that list was the most-rated games in the catalog. A probability model now scores the candidates first.
Reality
A dev.to developer prices each API request after the work by asking a second model how much the answer represents, and the first version threw away the calibration he was paying for by rounding its probabilities into three tiers.
Reality