> TL;DR: I ran 4 experiments using TypeSafe's Jev model to quantitatively measure geopolitical bias in AI recommendations. The results? 91.5% of the time, US models are placed…
ResourceResearch & data
Research & data
85 builds · page 2 of 3
I Tested TypeSafe’s Jev, a 478× Cheaper Decision Model, Against Claude, GPT-6, Kimi, MiniMax and DeepSeek. Only Two Beat It. | by Manjunath Janardhan | Sep, 2026 | Medium ... # I…
ResourceResearch & data
I Tested TypeSafe’s Jev, a 478× Cheaper Decision Model, Against Claude, GPT-6, Kimi,…
And then in September 2026, a well-funded frontier lab called TypeSafe AI (founded by Diogo Almeida, a co-inventor of ChatGPT at OpenAI) launched Jev. They proposed the exact…
ResourceResearch & data
I Built Non-Autoregressive Decision Models a Year Ago. Then a Frontier Lab Called It a…
Omi
Sponsored
Gentle daily affirmations on your Home Screen and Lock Screen.
TypeSafe describes Jev as a model that takes structured or unstructured state and returns typed, probabilistic decisions that software can use directly. ... ## Step 2: Understand…
ResourceResearch & data
How to Use Jev by TypeSafe AI: A Practical Guide to Building AI Decisions
Jev, a System One model from TypeSafe AI, evaluates supplied state against typed questions and returns choices, scores, and boolean probabilities without generating prose. The AI…
ResourceResearch & data
How to classify, route, and score with Jev and AI SDK
Actually Building With ... When TypeSafe AI came out of stealth in September 2026 with founder Diogo Almeida, a name attached to both ChatGPT and RLHF, the framing was almost…
ResourceResearch & data
Five Things Developers Are Actually Building With JEV (And Why It Matters)
Jev is the first public model from TypeSafe AI, released in early access on Sep 15, 2026. The announcement post is here. ... TypeSafe frames Jev as a new class of frontier model:…
ResourceResearch & data
Early experimentation using Jev to rethink harness UX - elvex
TypeSafe released Jev on September 15 as its first System One model. The interface is built around typed questions over some application state. Instead of generating text token…
ResourceResearch & data
Designing Production AI Systems with Jev | ArchCrux
▸ TypeSafe's Jev outputs decisions, not text (source: @aigc1024, @https1024). The model does only three things: boolean judgments ("search or not"), enum choices ("fast vs plan,…
ResourceResearch & data
AI Brief: Claude's Riemann Leap, Jev, Vercel's deepsec
typesafe.ai built Jev to answer yes/no and multiple-choice questions in under 500ms, not to write text. Here’s the architecture behind a model that gives up string generation to…
ResourceResearch & data
200x Faster, 400x Cheaper, Zero Words: Inside Jev
GitHubResearch & data
safer-with-jev
GitHubResearch & data
jevcal
Local decision cache keyed on (model, schema, state), with redaction and canonicalisation before hashing, for cheaper repeats and deterministic replay in CI.
GitHubResearch & data
jevcache
Chinese/English retrieval evaluation comparing Jev reranking with lexical, embedding, and fusion baselines, including judge-circularity analysis.
GitHubResearch & data
jev-search-rerank-eval
Retrieval and RAG: uses Jev Noul judgments to assess retrieved documents for relevance and usefulness as answer evidence, then sorts results and optionally filters them using a…
GitHubResearch & data
jev-reranker
Sitequiet
Sponsored
Uptime monitoring, status pages and client reports for agencies.
Fourteen-dataset reranking study with saved raw responses, paired bootstrap intervals, order-sensitivity checks, and no-relevant-document tests; comparisons remain study-specific.
GitHubResearch & data
jev-rerank-bench
Reproducible experiments on whether reranking with Jev improves a small RAG system, on a locked Turkish dataset, with quality, latency and cost reported together.
GitHubResearch & data
jev-rag-benchmark
Measures whether ORDER BY over a Jev probability is defensible (inversion rate, Score ordinality against a human grade, calibration, wording invariants, sort-key ties) under a…
GitHubResearch & data
jev-orderby-bench
Small Korean/English sample study with recorded responses, including medical-text questions; not a clinical validation.
GitHubResearch & data
jev-korean-benchmark
Independent Jev versus GPT-5.6 Terra comparison on three labeled classification tasks, reporting accuracy, calibration, latency, and cost.
GitHubResearch & data
jev-eval
GitHubResearch & data
jev-benchmarks
Independent synthetic-task study of Jev 1.13.0 framing sensitivity and failures, with raw responses and offline report checks.
GitHubResearch & data
jev-behavior-study
Compares a typed decision model with classical classification pipelines across eight datasets, with a published protocol and an interactive report.
GitHubResearch & data
Jev vs. ML
BAML language support for an AI if-statement: `.feels()` as a real, typed method backed by a decision model.
GitHubResearch & data
feelings
Cloudflare Workers has Jev now so I'm putting it to the test on https://t.co/1fK8HbSmPm - 7x faster search rerank compared to the current hybrid - 50x faster tagging of content vs GLM 4.7 Flash with no failures
GitHubResearch & data
jev-search
aeogeolabs
Sponsored
Get recommended in ChatGPT, Perplexity and Google AI Overviews.
Systematic-review data from trial reports, every answer a verbatim quote.
GitHubResearch & data
jev-reviewer
I made a DuckDB extension where you can use @typesafeai 's Jev to do quick classification of rows in any csv/parquet file or duckdb table about 10sec for 1k rows ~ better than using an LLM, way more ergonomic than a classifier game-changing for data analysis!
GitHubResearch & data
duckdb-jev
GitHubResearch & data
Jev Capability Atlas
JEV is insanely fast. We gave it a massive dataset based on thousands of outreach messages and asked: Which intent signals generated the most booked demos? 40 seconds later, we had the answer. Cost: less than $0.20. JEV can also rank leads, measure prospect-message fit, and
Using Jev to filter through my @wip todos It allows me to super quickly find all the instances where I increased revenue, got stuck, switched to a different SaaS provider, etc Things a regular keyword search would never catch
X postResearch & data
Filtering WIP todos by meaning
Jev can serve as better natural language search on websites. It can scan thousands of Zillow listings and classify properties by things you can't normally filter for - e.g. architecture, renovation status, proximity to freeways. This was done in <20 sec and costs $0.18 👇
JEV IS INSANE. I gave it 80 real exam questions and 297 practice ones. In 80 seconds, it told me which ones are most likely to appear on the real exam and which ones aren’t. All for $0.0256. Can't stop playing with @typesafeai 😁
Built a construction plan-set classifier with Jev. Proq turns civil and building plan sets into bills of materials using an LLM pipeline we built on GPT-4.1. Jev classified an entire 26-sheet plan set in 2.9 seconds for $0.0052. It matched GPT-4.1 and GPT-6 Astra on 100% of
Omi
Sponsored
Gentle daily affirmations on your Home Screen and Lock Screen.
