Jev

Of the six things people come to Jev for, only one works without caveats: checking the price. You cannot try it for free, cannot get a key today, and the Russian vendor calls Russian its second language.

Your code already asks an LLM “which of the three departments,” “spam or not,” “how angry is the customer.” Jev handles such questions in a third of a second for fractions of a cent; see the Norwegian documents example in chapter 2.

You need hundreds of thousands of small decisions an hour: moderation, sorting, scoring. Checking nine thousand product pairs cost $0,32 (paddo.dev, 19.09.2026).

You value the same answer to the same question. On public-procurement data, Jev repeated its decision in 98% of cases, while gpt-oss changed its answer on every sixth item (Habr, 21.09.2026).

You need text, code, or an answer based on an image. Jev accepts text only, and it guessed SVG drawings two and a half times less accurately than Claude Sonnet 5 using actual images (19.09.2026).

You want to ask one large question such as “can this code be deployed.” Almeida himself says this will work poorly: the model excels at small questions that combine into more complex tasks (17.09.2026).

You count, compare dates, or work in Russian without your own checks. The vendor says the model does not calculate and reads dates as text, while English is its primary language.

How to get a key and call the model, in chapter 3; how it differs from GPT-5 nano and Haiku, in the comparison. If you are just starting with AI, open the vibe coding introduction.

Three question types. Jev understands three question forms. Choice selects one item from a list, Score gives a score on a scale, and Noul answers “yes” or “no” with a number from 0 to 1. You can ask dozens of questions about one text at once; the model processes them in parallel, so response time barely increases.

Racing an LLM. In its launch video, one support ticket with three dozen questions was given to Jev and GPT-5.6 Terra. Jev answered in a tenth of a second; the LLM was still waiting for its first token. The vendor notes that short text favors its model (15.09.2026).

A judge for an agent. LangChain engineers Daniel Shi and Sean Roche ran the same weather-agent answers hundreds of times through four judges and checked them against human labels. Jev matched the human in every case, Claude Sonnet 4.6 was wrong one time in five, and the run cost 80 times less (our news story, 20.09.2026).

Norwegian hearings. Emil Lindfors gave Jev two dozen documents on resource rent in salmon farming and asked it to identify each author's position. The model got the position right five times in six and answered per document in a third of a second; DeepSeek without reasoning took almost three. More detailed question wording, to the author's surprise, made it less confident (lindfors.no, 18.09.2026).

Product catalog. On blog.r6i.it, product classification through a category tree was moved from an agentic GPT-5.2 loop to Jev. A product moved through the tree seven times faster, with half as many model calls. Paths matched seven times in ten, and a Jev error on the first branch is irreversible: it does not go back up the tree (21.09.2026).

Russian procurement codes. Vitaly Trenkenshu matched OKTRU codes to public-procurement items and asked models to select the best code from a list of candidates. Jev hit the right code more often than any other, including Gemini 3.1 Pro, which cost 60 times more. English is the model's primary language, and the author recommends testing on your own data (Habr, 21.09.2026).

Where it stumbled. Bartosz Mikulski converted four hundred Quick, Draw! sketches into SVG text and asked it to guess the object. Jev guessed one in three and called more than half the sketches airplanes; Claude Sonnet 5, using the actual images, guessed nine out of ten. The model cannot see images at all (19.09.2026).

More. On Hacker News, a user named silbercue guided Jev through websites as a browser agent: the model chose where to click and made no mistakes in six scenarios, while a separate small model typed text into fields.

Our experience. We asked Jev to re-judge two hundred recent posts from our news feed and checked it against our large-model judge's decisions. They matched three times in four, and from Russia via OpenRouter a response arrived in 0,42 s; the entire run cost three cents. Where Jev is unsure, the judge is usually uncertain too; details in the article (21.09.2026).

Waitlist. Registration at TypeSafe is closed: on typesafe.ai you submit an application, and a key is issued after an invitation in the console.typesafe.ai console. The documentation lists no free tier or starter credits.

Call. One endpoint, POST https://api.typesafe.ai/v1/systemone, a key header, and a two-field body. state holds the text; questions holds a list of questions with options. SDKs are available for Python (typesafe_sdk) and JavaScript (@typesafe-ai/sdk); the default model is jev-latest, currently jev-1.13.0.

Through gateways. Without a waitlist, the model is available through Vercel AI Gateway as typesafe-ai/jev, and through Cloudflare AI and OpenRouter as typesafe/jev-1.13, at the same price. At Vercel, this is not a regular chat call but a separate evaluate function in AI SDK 7, because Jev has its own request format.

For agents. TypeSafe has released a skill for coding agents. In Claude Code, install it with claude plugin install typesafe@typesafe-ai; it teaches the agent how to write questions for the model.

Limits. 64 thousand tokens per request and 32 thousand for the text plus the longest question, 1 200 requests per minute, and 250 thousand tokens per second. The vendor warns that limits fluctuate and may change without notice while demand exceeds capacity.

From Russia. The working route is the same as for other foreign APIs: overseas payment and a gateway that already has an account. TypeSafe does not name supported countries, cards, or phone numbers in its public docs, so availability through them cannot be checked. The intermediary polza.ai promises ruble payments; that is its claim and has not been tested with a purchase.

Vendor measurement. On its four real-world tasks, TypeSafe measured Jev at 67,8% accuracy for $0,0004 and 0,4 s per case. GPT-5.6 Sol scored 74,1%, Claude Opus 5 73,1%, GPT-5.6 Terra 67,9% at $0,03 and 10 s. The benchmark for correct answers was the averaged responses of GPT-6 Astra and Claude Fable 5.1, rather than human labels (evals.typesafe.ai, 21.09.2026).

Where 193,6 and 444,6 come from. The vendor derives these multipliers on its homepage from the same measurement and calls them an upper bound. By our recalculation of its table, 193,6 comes from dividing Claude Sonnet 5's 78,1 s by Jev's 0,4 s, and 444,6 from dividing Opus 5's $0,176 by $0,0004. Speed is compared with the slowest model, cost with the most expensive.

Independent runs. Three measurements from 18–20 September agree: versus low-cost LLMs, the speed gain is two to three times. On Banking77 with 77 classes, Jev scored 83,2%, GPT-5.4 nano 79,3%, GPT-5.6 Terra 87,5%, and a small classifier trained on these data 93,3% (ickma2311, 18.09.2026). On the full Banking77 test of 3 080 examples, JevBench got 80,3%, and on a phishing set measured 239 ms latency from France versus 687 ms for Claude Haiku 4.5 (20.09.2026).

It depends on the question. On phishing emails, Jev with one question scored 62,6% versus 81,3% for Haiku 4.5, but after splitting it into five questions with tuned weights, 95,0% (XenoSpectrum, verified 21.09.2026). A simple regular expression scores 91,8% on the same set.

Russian data. Vitaly Trenkenshu asked models to select the best OKTRU code from a candidate list for 928 public-procurement items; the directory contains 77 772 codes. Jev achieved 0,775 accuracy for $0,56, Gemini 3.1 Pro 0,753 for $33,46, gpt-oss-120b 0,741 for $1,10 (Habr, 21.09.2026).

Token price. $0,042 per million input tokens; output tokens are free. In its documentation, the vendor counts in billions: $42 per billion. There is no separate pricing page, plans and minimum payment are not stated, and the price appears only in the model description (checked 21.09.2026).

Task price. A thousand hearing documents cost $0,22 versus $1,31 for DeepSeek without reasoning (lindfors.no, 18.09.2026). Classifying 1 018 articles cost $0,08 (Habr, 19.09.2026). A Doom bot made 10 requests per second for $7 an hour; this is the vendor’s only real-world figure.

The price catch. In its blog, TypeSafe writes that it cannot prove the absence of a subsidy and expects a long period to confirm the price is sustainable. Its homepage, meanwhile, says the model is served profitably. Limits change without notice.

From Russia. TypeSafe hides payment behind login, and its documentation says nothing about countries or cards. Payment methods for foreign APIs are collected on a separate page.

Format, not truth. Jev cannot return an option outside your list, so the “zero hallucinations” claim on the homepage is true in that sense. It can choose the wrong option from the list; the vendor says so directly: the model guarantees the response format, not the correctness of every decision.

What confidence means. Calibration is promised across answer groups: of decisions with a 0,8 probability, around 80% should be correct. It guarantees nothing about an individual answer. In ickma2311’s run, Jev’s confidence did not separate its errors better than a standard LLM’s verbal confidence (18.09.2026). The vendor has not published a paper, calibration curve, or calibration-error figures; our analysis of this claim appeared in the feed on 21.09.2026.

Nine weak spots. TypeSafe maintains its own failure page for jev-1.13: literal reading of the question, arithmetic, date comparisons, references across several steps, long text with noise, adversarial insertions, conflicting criteria, rules that must align across answers, and generation. The vendor has one recommendation: calculate and compare in code, leaving judgment to the model (checked 17.09.2026).

Data. According to the documentation, the model is not trained on customer requests, and enterprise customers have a no-data-retention mode. Jev cannot be fine-tuned on your data; the same weights serve every account.

Russian language. English is named as the primary training language; other languages are “processed, but not equally well.” In the only measurement on Russian data, public procurement listings on Habr, Jev won, but the author makes the same caveat. Test it on your own texts and watch the confidence.

Who is behind it. Diogo Almeida is the fourth of twenty authors of InstructGPT, the 2022 paper that led to ChatGPT, and one of around fifty contributors to the ChatGPT launch blog post. The phrase “co-inventor of ChatGPT” from his tweet does not appear in OpenAI documents. The company was founded in San Francisco in 2024 and received $40 million from DCVC on 15.09.2026.

Jev or GPT-5 nano. Nano can write text and costs $0,05 for input and $0,40 for output. On Banking77, Jev was four points more accurate and twice as fast per call, 0,43 s versus 0,9 s (ickma2311, 18.09.2026). If you need one decision and no text, Jev is cheaper; if you also need to write a customer reply, nano handles both tasks.

Jev or Claude Haiku 4.5. Haiku understands images, writes text, and costs $1 for input. On phishing with one question, Haiku scored 81,3% versus Jev’s 62,6%; after breaking it down into five questions, Jev reached 95,0% (XenoSpectrum, checked 21.09.2026). Jev’s win here comes at the cost of manual work on the questions.

Jev or your own classifier. A small model trained on the same Banking77 data scored 93,3% versus Jev’s 83,2% and responded in 0,01 s (ickma2311, 18.09.2026). But such a model must be trained for every task, while Jev takes the task from the question text without training. On Hacker News, it was called more of a classifier than a language model, and the founder replied “very accurately.”

Where each call fits. Jev does not fit the usual chat format, so it cannot be plugged into OpenAI-compatible libraries with a single line. Nano and Haiku work through the familiar OpenAI and Anthropic APIs.

Three days passed from leaving stealth to the first independent measurements, six to the first measurement on Russian data. Everything below is dated by the authors' publications.