Jev is TypeSafe AI's debut System One Model, in waitlisted early access since September 15, 2026, with a 64,000-token context window and no public SWE-bench, GPQA, or MMLU scores. It replaces open-ended text generation with a fixed, developer-declared set of typed decisions, aimed at teams automating classification, routing, and scoring inside existing software rather than building a chatbot.
Jev is TypeSafe AI's first System One Model, launched in September 2026. Instead of generating text token by token, it takes a state plus typed questions and returns calibrated decisions with confidence scores in 70 to 500 milliseconds, using a non-autoregressive parallel sampler rather than sequential token generation.
Where it sits
- $0.032/M$ per 1M tokensBlended price (3:1)Lower is better#4 / 62peer median $1.70/Mvendor price, checked by HokAI
- --tokens/sOutput speedHigher is better-- / 37peer median 90 tok/scited: Artificial Analysis
- --% solvedSWE-bench VerifiedHigher is better-- / 28peer median 78.3%per source, see benchmark scores
- --% correctGPQA DiamondHigher is better-- / 44peer median 88.3%per source, see benchmark scores
Ranks are against GA models on HokAI that publish the same figure; ties share a rank.
Provider: TypeSafe AI · Family: System One Models
More about TypeSafe AI on HokAI
Context window: 64,000 tokens
Input modalities: text · Output: text
About Jev
Jev is TypeSafe AI's first public model and the debut of what the company calls a System One Model: a new class of frontier model built to return fast, structured decisions rather than generated text. TypeSafe announced Jev on September 15, 2026, in a launch post from founder Diogo Almeida, who said the model runs on a new model architecture, a parallel sampler for efficiency, and a training method the company calls Reinforcement Learning for Calibrated Decisions (RLCD). TypeSafe has not disclosed Jev's parameter count or further architectural detail. The hosted API is versioned; TypeSafe's own documentation lists jev-1.13.0, jev-latest, and jev-preview as selectable model identifiers.
Unlike a conventional LLM, Jev does not generate text token by token. A single call sends a state (unstructured text or a structured JSON object describing a program's current situation) plus one or more typed questions, and Jev returns every answer from one parallel pass instead of sampling sequentially. TypeSafe documents three question kinds: noul (a yes/no question returning the probability of yes), choice (pick one option from a declared set, capped at 255 options, returning probabilities and a confidence score), and score (rate the state against an ordered list of descriptive levels, returning a continuous value with probabilities and confidence). Because the space of possible outputs is declared by the developer up front, TypeSafe says the model cannot return a value outside that schema, a guarantee about output structure rather than about the correctness of the underlying judgment.
Jev accepts up to 64,000 tokens per request, with the state portion plus the longest single question capped at 32,000 tokens, per TypeSafe's published API documentation. Input is text only: a string, a JSON object, or an array of text values. TypeSafe's own demos note that non-text data, such as a game screen, has to be pre-processed into a text or structured description first, since Jev does not accept images, audio, or video directly. Documented account-level rate limits are 250,000 tokens per second and 1,200 requests per minute, though TypeSafe describes these as adjusting dynamically during the early-access period rather than a fixed guarantee.
TypeSafe has not published Jev results on standard public benchmarks such as SWE-bench, GPQA, or MMLU, saying the model was not built for the open-ended reasoning or writing tasks those measure. Instead the company built its own workflow evals: a small set of decision workflows scored against a reference answer it computes as the average of two other frontier models. TypeSafe's launch post states that its widely repeated home-page claims of being 193.6 times faster and 444.6 times cheaper on these tasks come directly from that evaluation, while cautioning in the same post that the workflows were designed by its own capabilities team, that the reference answers are biased toward OpenAI and Anthropic's models, and that it cannot yet prove its pricing is not subsidized. Separately, TypeSafe states Jev's end-to-end response time as 70 to 500 milliseconds, against a stated 3 to 329 seconds for the frontier LLMs it compares itself against on similar tasks.
Jev bills only for input tokens, with output free of charge, which the company calls too cheap to meter; exact figures are in the pricing FAQ below. TypeSafe says this undercuts typical frontier-LLM input pricing by roughly an order of magnitude, and that those LLMs typically charge several times more for output than for input, a cost shape Jev does not share since it has no output charge at all. There is no published free tier, and access requires an approved early-access account.
Jev is reachable only through TypeSafe's own hosted API while the product is in a waitlisted early-access period; there is no self-hosting option and no downloadable model weights. TypeSafe has not listed availability through AWS Bedrock, Google Vertex, or Azure. Vercel has published a community guide showing how to call Jev from its AI SDK, so far the only third-party framework integration documented publicly.
TypeSafe positions Jev for what it calls AI-powered workflows: classification, routing, scoring, extraction, and branching logic embedded inside existing software, where a conventional LLM's free-form text output would need parsing and validation and could still land outside the expected shape. Its own public demos, a Doom-playing bot driven by a text description of game state and a Wikipedia-link-following Wikiracing bot, are built to show the model choosing among many discrete options quickly and cheaply rather than holding a conversation. TypeSafe explicitly does not position Jev for human-facing chat, open-ended writing, or general coding, jobs it says still belong to conventional LLMs with a person supervising the output.
TypeSafe has not disclosed a system card, a training data cutoff date, or a data retention policy for Jev, and has not stated whether input sent to the early-access API is used for further training. The company has not published SOC 2, ISO 27001, HIPAA, or other third-party compliance certification for the product.
Pricing
Input tokens: $0.042 per million ($42 per billion). Output tokens: free, which TypeSafe calls too cheap to meter. No published free tier; access requires an approved early-access account, and TypeSafe says its rate limits adjust dynamically during this period.
What a real job costs
| Job | Input | Output | Total |
|---|---|---|---|
| Summarise a 20-page PDF | $0.0013 | $0 | $0.0013 |
| Support reply | $0.0001 | $0 | $0.0001 |
| One coding agent run | $0.0084 | $0 | $0.0084 |
Budgets: 20-page PDF = 30k in / 1k out · Support reply = 2k in / 300 out · Coding agent run = 200k in / 20k out. Computed from the vendor's per-token prices at render time; cached-input discounts are not applied.
Key Features
- Typed decision primitives: Three typed question kinds (noul, choice, score) let a single call return a yes/no probability, a pick from up to 255 declared options, or a continuous rating, always inside a developer-declared schema.
- Parallel, non-autoregressive sampling: Jev returns every requested answer from one parallel pass instead of generating tokens one at a time, which TypeSafe credits for its sub-second, real-time response speed.
- Calibrated confidence on every answer: Every choice and score response carries a confidence value between 0 and 1 that TypeSafe derives from the shape of the output probability distribution, meant to flag when a decision needs human review.
- Input-only pricing: TypeSafe charges only for input tokens and gives output away free, a cost shape unlike standard per-token chat model pricing (see the pricing FAQ for exact rates).
Pros
- TypeSafe reports sub-second, real-time response times fast enough for the live game and robotics demos the company has shown publicly.
- Charges nothing at all for output tokens, unlike standard per-token chat models that typically price output well above input.
- Output is constrained to a developer-declared schema (noul, choice, or score), which TypeSafe says makes a schema-violating response mathematically impossible.
Cons
- Accepts text only; images, audio, and video must be converted to a text or JSON description first, so it cannot process raw screenshots or audio directly.
- TypeSafe has not released independent SWE-bench, GPQA, or MMLU numbers, so how Jev's judgment holds up beyond the company's own workflow evals is still an open question.
- Only reachable through a single waitlisted early-access API; there is no way to run it yourself or reach it via a major cloud AI marketplace today.
Frequently Asked Questions
What does Jev cost per million tokens?
Input runs $0.042 per million tokens ($42 per billion), and TypeSafe charges nothing for output, calling it too cheap to meter. TypeSafe contrasts that with the $0.20 to $10 per million input tokens it says typical frontier LLMs charge, though it says it cannot yet prove its own pricing is not subsidized.
How does Jev's performance compare with models like GPT-5.6 or Claude Opus 5?
TypeSafe has not published Jev scores on standard benchmarks such as SWE-bench, GPQA, or MMLU, arguing those measure tasks the model was not built for. On its own internal workflow evals, TypeSafe says Jev matches the intelligence of an average of two larger frontier models while answering far quicker, a comparison it built and has not had independently audited.
Is Jev open source, and can you self-host it?
No. Jev is proprietary and available only through TypeSafe's hosted API during a waitlisted early-access period, with no model weights, detailed architecture, or self-hosting option released.
Does TypeSafe train on the data sent to Jev?
TypeSafe has not published a retention or training-data policy for Jev's early-access API and has not stated whether submitted state or questions are used for further model training. Teams evaluating it for sensitive workloads should confirm the current policy with TypeSafe directly before sending production data.
Who should use Jev, and who should look elsewhere?
Jev suits teams that need fast, cheap, schema-constrained decisions embedded in existing code, such as classification, routing, or scoring steps, rather than open-ended writing or conversation. Anyone needing multimodal input, general reasoning, code generation, or a chat interface should reach for a conventional LLM instead, since Jev takes text only and hands back typed decisions rather than prose.