TypeSafe Launches Jev, a Decision-Only Model That Never Writes a Word

TypeSafe (founded by Diogo Almeida, one of the researchers behind ChatGPT/InstructGPT) has released its first model, Jev, positioned not as a better chatbot but as a "decision engine": given an application state plus a set of typed questions, it directly outputs options, scores, and probabilities for code to execute, without generating any text. The company claims it is 20-200x faster and 40-400x cheaper than LLMs (output tokens are free; input pricing works out to about $0.042 per million tokens), with latency of 70-500 milliseconds, and uses a new training method called RLCD. In a demo, Jev played Subway Surfers in real time, showcasing real-time decision-making.

Confirmed

• Jev takes a "System One" approach: instead of generating text token by token, it directly selects a decision from a fixed set of candidate answers with a confidence score, in contrast to the LLM's token-by-token "System Two" generation
• Latency of 70-500ms, free output tokens, positioned as a third path beyond hybrid classifiers and LLMs
• Early hands-on from independent AI engineer @mostlyidempotent confirmed it returns structured decisions rather than text, with latency around 70ms
• @mikegiannulis ran an offline replay test (real historical inputs + manually labeled ground truth, with pass/fail rules preset in advance): for manuscript material reranking, embedding similarity ranked the correct passage first 60% of the time, while Jev raised it to 90%; portal intent routing scored 90.5% vs. 83.8% for the original keyword routing; "I've already done this" detection hit 98.3% vs. 96.6% for the regex approach; 95% accuracy and Brier 0.057 on 148 faithful/fabricated pairs; the replay also caught a customer service message the old classifier had misjudged in production to the team's embarrassment (e.g., "please text, don't call" was classified as other)

Unconfirmed

• "20-200x faster and 40-400x cheaper" are official claims; independent benchmarking remains insufficient; the founder's identity also carries传播 noise—most reposts say "ChatGPT co-inventor," while @adrianscottcom noted Diogo Almeida is an InstructGPT co-author; the exact attribution varies
• Small-scale replays can't prove perfect calibration in production, and @mikegiannulis himself explicitly remained measured

Why it matters

• Jev represents a "small model fills the gap" paradigm: not replacing LLMs, but replacing the roundabout practice of "have an LLM write a paragraph then dig out the answer" with probabilistic outputs for deterministic yes/no, routing, and classification tasks
• @mikegiannulis also disclosed a failure list: their own LLM won about 2:1 on subtly-worded author emails, it couldn't help with subjective judgments like "this seems too salesy," and showed no improvement on data-sparse sales lead scoring—blind spots like these can't be found on AI leaderboards, only exposed by replaying your own data, providing a rare honest reference for the model's applicability boundaries

2026-09-17 ~ 2026-09-18 · 44 related posts

Full story(9 episodes)→

• Episode 1: TypeSafe exits stealth with decision model Jev and RLCD training method(2026-09-16, 86 posts)
• Episode 2: TypeSafe AI launches Jev, a dedicated evaluation model showing major speed and cost gains in tests(2026-09-16, 8 posts)
• Episode 3: Vercel fx to adopt Jev safety reviewer, up to 18x faster(2026-09-17, 3 posts)
• Episode 4: TypesafeAI Launches Zero-Shot Classifier Jev on AI Gateway(2026-09-17, 2 posts)
• Episode 5: TypeSafe Launches Jev, a Decision-Only Model That Never Writes a Word(2026-09-17, 44 posts)
• Episode 6: TypeSafe AI's Evaluation Model Jev Lands on Vercel and Cloudflare Gateways(2026-09-17, 4 posts)
• Episode 7: Typesafe AI's New Model Jev Delivers Fast, Cheap Classification(2026-09-17, 2 posts)
• Episode 8: Jev Classifies Tweets 6x Faster and 40x Cheaper than Fastest LLM(2026-09-17, 2 posts)
• Episode 9: Jev Outputs Structured Decisions Directly at $0.042/M Tokens(2026-09-18, 2 posts)

• Jev Debate: Engineers Forget Encoder-Only Classifiers Have Existed for Years — brandon_galang · 2026-09-17
• Jev buzz reveals many don't know encoder-only classifiers have existed for years — RichmanRonald · 2026-09-17
• Six real-world uses for TypeSafe's Jev judgment model, 4x faster than Gemini in evals — HamelHusain · 2026-09-18
• Six Real Uses for Jev, the Judgment Model: From Fact-Checking Scripts to Debugging Agents — HamelHusain · 2026-09-18
• Self-Proclaimed ChatGPT Co-Inventor Launches Jev, Claims 200x Speed at 1/400 Cost — iamrobotbear · 2026-09-18
• Jev hype reveals two things: weak builder skills and the case for renting inference — edgarpavlovsky · 2026-09-18
• [source] TypeSafe, founded by a ChatGPT researcher, launches Jev — a decision engine that can't write — Scobleizer · 2026-09-18
• ChatGPT co-inventor launches Jev model with AI detector API 100x cheaper than Pangram — gaganghotra_ · 2026-09-18
• Jev Ditches Autoregression: A Model That Only Outputs Structured Decisions — karminski3 · 2026-09-18
• Jev: A Non-Autoregressive Decision Model That Outputs Structured JSON in 70ms — karminski3 · 2026-09-18
• Jev, the non-chatbot model, returns decisions in 70-500ms at $0.042/M input tokens, free output — adrianscottcom · 2026-09-18
• Jev and System One Models: fast decision-only AI that outputs typed answers, not text — blaizedsouza · 2026-09-18
• Dev: Jev could run cheaply in browsers — classification tasks were shoehorned into LLMs — GabGarrett · 2026-09-18
• TypeSafe launches Jev, a "System One" model that answers typed judgments in ~100ms — aronchick · 2026-09-18
• Jev, a new 'decision' model from a ChatGPT co-creator, plays Subway Surfers in real time — Cagnazzo82 · 2026-09-18
• ThursdAI Podcast Dives Into Jev Real-Time Models and Computer-Use Agents — altryne · 2026-09-18
• ThursdAI: TypeSafe's non-LLM Jev hailed as a ChatGPT moment, Dario proposes three ways to pace frontier AI — altryne · 2026-09-18
• A day with Jev: a fast classifier, best suited for Computer Use and game AI — jiayuan_jy · 2026-09-18
• After a day with Jev: a blazing-fast classifier, not a GPT replacement — jiayuan_jy · 2026-09-18
• Jev Launches as a Decision-Only Model: 20-200x Faster, 40-400x Cheaper Than LLMs — thisiskp_ · 2026-09-18
• [source] TypeSafe Jev returns typed decisions, not text: 70ms latency, free output tokens — mostly_idempotent · 2026-09-18
• A 20-cent AI that can't write a sentence caught 13 requests our classifier missed — mikegiannulis · 2026-09-18
• Replaying 3,300 Real Production Decisions to Test Small Model Jev — mikegiannulis · 2026-09-18
• Before buying a smarter model, check you asked the wrong job: Jev test wrap-up and rollout rules — mikegiannulis · 2026-09-18
• [source] Same Material, Better Order: Reranking Lifts Correct-Passage Hit Rate to 90% — mikegiannulis · 2026-09-18
• 13 'Please Text Me, No Calls' Messages Misrouted — Small Model Caught All of Them — mikegiannulis · 2026-09-18
• Intent Routing 90.5% vs 83.8%, 'Already Did That' Detection 98.3% vs 96.6% — mikegiannulis · 2026-09-18
• 95% Accuracy, 0.057 Brier Score on Faithful-vs-Fabricated Pairs — With Caveats — mikegiannulis · 2026-09-18
• What Didn't Work: LLM Won Nuanced Emails 2:1, Three Tasks Where Small Model Failed — mikegiannulis · 2026-09-18
• The 20-Cent Bill Is a Headline — Real Wins Are 0.2s vs 2–4s Latency and Fewer Mistakes — mikegiannulis · 2026-09-18
• Offline Wins Earn a Shadow-Mode Test, Not the Keys to the Whole Business — mikegiannulis · 2026-09-18
• Jev, a decisions-only model, matches gpt-4o-mini 50/50 while being 2.4x faster and 3.4x cheaper — Al_Grigor · 2026-09-18
• Jev, the model that isn't an LLM: $42 per billion input tokens, returns probabilities — thursdai_pod · 2026-09-18
• Computer-use agents: 273 tests across 23 assistants, Muse books a haircut by phone — thursdai_pod · 2026-09-18
• ThursdAI: TypeSafe's Jev decision model hits 70-500ms at $42/1B input tokens, outputs free — thursdai_pod · 2026-09-18

9 near-duplicate retellings: threepointone · hardimanjames · HankYeomans · multiply_matrix · multiply_matrix · multiply_matrix · GabGarrett · multiply_matrix · hardimanjames