Jev Press

TechCrunch reports TypeSafe AI raised $870 million at a $7.5 billion valuation, led by Andreessen Horowitz with Sequoia and DCVC, weeks after Jev's September 15 release. The startup claims a third of Fortune 500 companies already use the model.

Per WSJ's search excerpt, the three-week-old Jev has set Silicon Valley abuzz with talk of alternatives to large language models and is already sparking copycats. It classifies inputs to predetermined outputs: yes or no, a numerical score, or answers on a list.

Fortune reports Jev went viral among developers, with nearly 13% of Vercel's paid teams trying it within a day of its gateway arrival. Three weeks after Jev's launch, OpenAI rolled out a competing Decisions API built on GPT-6 Luna.

InfoWorld covers a HiddenLayer report finding Jev more vulnerable than traditional LLMs to adversarial text and prompt injection, with analysts calling the methodology sound but debating whether flaw reports this early are premature.

Bloomberg Law reports TypeSafe AI raised about $870M at a $7.5B valuation in a round led by Andreessen Horowitz, weeks after the Jev launch video went viral. The startup says a third of Fortune 500 companies use Jev but declined to name any.

SiliconANGLE reports TypeSafe closed $870M at a $7.5B valuation and explains the model's pitch: three request types, yes/no answers, picking from a list, or a score, return structured output and a confidence number instead of text.

The Economic Times profiles Diogo Almeida, who left OpenAI in 2024 and spent about two years developing the idea behind TypeSafe AI. It cites WSJ on the roughly 25% Fortune 500 adoption and a $40 million DCVC raise.

The Register covers HiddenLayer's tests of Jev and a decision-tuned Qwen model: a planted ad stole a browser agent's click on a simulated store page, and tokenbreak edits snuck toxic prompts past verification. TypeSafe did not immediately respond.

Adrian Bridgwater's September 21 report on TypeSafe's launch pitch: a model that does not write but decides, a new architecture and RLCD training, and Vercel's Jev adoption numbers through its AI Gateway.

Andreessen Horowitz announces it is leading the investment in TypeSafe AI, calling Jev the fastest growing model it has seen at 1 trillion tokens generated in 3 days. The firm claims 1/100 to 1/500 frontier-model cost and 100x speed for classification, and says 25% of Fortune 500 enterprises have integrated Jev.

HiddenLayer researchers tested Jev and an in-house System One-style model: on a simulated checkout page an attacker-controlled ad made the browser agent click away from the purchase, and TokenBreak-style edits lowered toxicity scores for prompts both models had flagged.

HiddenLayer's insight post walks security teams through its Jev tests: on a simulated checkout page Jev gave the attacker ad's button 91% against 9% for Place order, and TokenBreak edits cut one prompt's toxicity score by more than two thirds so Jev stopped flagging it.

Pillar Security analyzes Jev and System One models: a smart-home demo answered three typed questions in about 200 ms with no language model in the loop, while prompt scanning, output filtering and chain-of-thought monitoring all have nothing to inspect. TypeSafe's docs say state is not treated as hostile by default.

Rogue Security's post uses Jev and System One models to frame agentic security as a decision problem: the same send_file action should be allowed, blocked or reviewed depending on who requested it and where the data goes, illustrated with policy case files rather than product results.

The Injection's brief on HiddenLayer's tests: a browser agent powered by jev-ultrafast followed instructions planted in a web page and picked an attacker's ad over the real checkout button, and TokenBreak edits let toxic prompts slip past Jev's scoring. The post records no fix or response from TypeSafe.

Ian Maurer's ThinkThen analysis argues Jev's faster and cheaper are measurable today while its better is a bet on RLCD: unlike RLHF, which pleases humans, and RLVR, which optimizes benchmarks, Jev is trained to be reliable for programmatic use, in Diogo Almeida's own framing.

Same-day reaction to Jev from four engineers: Mario Zechner called it basically a general purpose classifier, Thorsten Ball praised the onboarding, Malte Ubl said it reached AI Gateway, and Armin Ronacher expects fast followers. Wes Bos measured $0.000232 a call and there were no evals yet.

A podcast episode on Jev as a decision model rather than a chat model: hosts Luis Calderon and Carina discuss roughly 20x faster Coach testing, why that preliminary figure differs from TypeSafe's vendor benchmarks, and why type safety does not guarantee a correct decision.

A BIG CHANGE analysis of the Diogo Almeida Latent Space interview: Jev's most consequential proposition is the unit of automation, a bounded judgment inside an existing process, and the launch claim that Jev can't hallucinate needs a narrow reading because schema matching does not establish truth.

Simon Willison released llm-typesafe 0.1a0, an LLM plugin for TypeSafe's Jev: install with llm install llm-typesafe, set a typesafe API key, then ask noul yes/no questions, Choice questions and Score questions from the command line. A demo refund question returned noul 0.99.

Bloomberg News reports Jev went viral with the promise of a cheaper, faster alternative to large language models, its launch video nearing 40 million views on X. Per the Financial Times, investors have offered funding valuing TypeSafe AI above $10 billion.

Business Insider's Stephen Council reports from a Jev hackathon at CodeRabbit's office, where developers said they had woven the model into agents and search tools. The piece cites nearly 40 million launch video views and a Discord server past 100,000 members.

VentureBeat's Sam Witteveen argues a large share of enterprise AI only needs a label, not generated text. Per TypeSafe, Jev returns choices, scores and yes/no answers in 70 to 500 ms at $0.042 per million input tokens, with output tokens free.

VentureBeat's Louis Columbus reports Jev is spreading into agent infrastructure faster than enterprises can audit it. In one Octomind test a planted field cut a block probability from 0.76 to 0.48, and TypeSafe and Pydantic both warn inputs can steer answers.

Forbes' Rashi Shrivastava profiles TypeSafe's $40 million DCVC-led seed at a $200 million valuation, and CEO Diogo Almeida's argument that software needs models reporting calibrated confidence rather than confident-sounding paragraphs.

TNW reports Jev reached nearly 13% of Vercel's paid teams within 24 hours, the AI Gateway's fastest adoption: twice the GPT-5.6 family's share and more than 6x Fable 5.1's, at $0.042 per million input tokens with output unmetered.

Vega's proof of concept put Jev in front of its alert triage agent on two tenants: it gated off 15% of alerts on the busy tenant and a third on the quiet one, about 98% correct by three LLM judges verified by its own analysts.

TWIML episode 779 features TypeSafe co-founder and CEO Diogo Almeida on Jev, the company's model for bringing fast, reliable intelligence directly into software, and why text-generating models are poorly suited to many real-world automation decisions.

On The MAD Podcast, Matt Turck talks with TypeSafe CEO Diogo Almeida about why intelligence per dollar matters more than enterprise adoption, and why RLVR benchmark gains still leave models 95% Einstein, 5% Mr. Bean.

In Law://WhatsNext's return episode, guest Chris Bridges of AI-native law firm Tacit Legal demonstrates Jev returning structured contract answers in 0.3 seconds where a chat model took 9.6 seconds.

The Rundown's news brief reports TypeSafe's September 15 early access launch of Jev, a model that answers preset questions inside software, and flags that the company's can't-hallucinate guarantee concerns format, not correctness.

DigitalToday's Korea report relays The Information's September 24 story: TypesafeAI is seeking $1 billion, with some unnamed investors proposing valuations above $10 billion, while the talks remain early and terms could change.

An analysis of TypeSafe's $870 million Series A at a $7.5 billion valuation that corrects the 24-days framing, showing the $40 million seed was priced more than a year earlier, and notes the close came in below September's $10 billion talk.

INXEN's syndication of Financial Times reporting says TypeSafe, last valued at $200 million, is fielding offers that would value it at $10 billion or more days after launching Jev, with Vercel citing record paid-developer interest.

Braintrust's provider page covers two ways to use TypeSafe Jev: instrumenting System One calls as decision spans, and running Jev as an LLM judge that turns typed choices into scores or labels.

Langfuse's guide to decision-model evaluators runs OpenAI's gpt-6-luna or TypeSafe Jev over observations, where each Choice, Score or Yes/no question becomes one Langfuse score.

Langfuse's notebook integration traces TypeSafe Jev System One calls through the OpenInference instrumentor, which patches the Python client in place and emits OpenTelemetry spans.

Langfuse's September 18 post presents Jev as a decision model for routing, classification and eval verdicts, and walks through scoring production traces with Jev-as-a-judge in Langfuse.

LangChain's September 25 post argues Jev's typed decisions fit LangGraph's deterministic orchestration, with a litigation review example and a Stagehand browser case where act() latency fell to 0.46 seconds.

A builder's walkthrough of System One, a dual-process browser agent that resolves routine steps with four parallel Jev questions in 70-150ms and falls back to Gemini or Claude below 0.55 confidence.

Tom Joseph's analysis walks through Jev's typed primitives and finds three claims that survive scrutiny: the parallel sampler, the price and speed, and the RLCD training whose calibration plot is unpublished.

Octomind 0.54 hands eight supervisor judgments to Jev, from risky tool-call checks to skill routing and done verification, reporting 0.5-1.3 second round trips at $0.00001-$0.0001 per decision.

A Spanish-language analysis of Jev's five-day rise to critical infrastructure, the documented injection finding that moved a block verdict, and the six controls it says to add before production.

A community roundup of every public Jev evaluation: TypeSafe's own workflow evals plus at least twenty independent builder benchmarks. The site reports speed medians of 127-422 ms per call leading every LLM compared, while accuracy splits by task.

Blackwell Systems' adversarial technical due diligence on Jev, built only from public material and live API calls. The author reproduces TypeSafe's published addressee numbers within a point and finds a real but noise-type-specific robustness moat.

An independent hands-on evaluation that ran Jev live on 16,379 benchmark requests and reports it is no frontier model but genuinely useful for cheap real-time sub-frontier judgement, with 82.7% MMLU-Pro and 76.5% GPQA Diamond.

A LessWrong pilot testing Jev as a trusted monitor in the ControlArena APPS backdoor setting: one yes/no question reaches AUROC 0.976 and catches 90% of backdoors at a 2% false-positive rate for about $0.04 per thousand submissions.

A playground operator's audit of every published Jev number: TypeSafe's four-workflow evals, the favourable homepage demo, independent tests and its own calls. It concludes the price and speed are real and the intelligence claim is roughly mid-tier frontier on decision workflows.

Celeris's launch post for Celeris-1-decision, a hosted diffusion language model for typed decisions. It reports 80.0% on the full jev-bench, the lowest Brier scores of five decision APIs tested, a 67 ms median and $0.04 per million input tokens.

AICrier's news-feed summary of Celeris-1 Decision: a multimodal diffusion model for bounded structured decisions that Celeris says reaches 80% on JevBench, 7 points above Jev 1.13.0, with 67 ms median latency.

Interactive comparison of 21 decision models, with direct-provider and OpenRouter routes measured separately on 1,100 public jev-bench items. The charts plot accuracy against benchmark cost, official prices and median latency, with Jev 1.13 and 1.14 measured separately.

Doug Does' benchmark dashboard for JevBench: 534 questions, 48 ranked models, composite scores that already include cost, head-to-head comparison, latency and a Pareto frontier. Jev 1.13.0 tops the table at 74.40.

YourStory's introduction to Jev, the transformer-based model that returns structured decisions with probabilities instead of paragraphs or code, covering the September 15 stealth exit, the $40M DCVC-led seed, and TypeSafe's pricing claims.

AIMag.no's week-in-review of Jev's waitlist-to-open-access sprint, Vercel and LangChain integrations, and Kognitos and GPTBots.ai partnerships, contrasting TypeSafe's 193.6x speed claim with a roughly 7x user measurement and one prompt-injection test.

ThursdAI's September 17 episode covers TypeSafe's Jev launch with founding DevRel Allie Laabs: choice, score and null primitives, $42-per-billion input pricing, and a live X timeline classifier running at 12 tweets a second.

An episode of The Merge (by CodeRabbit) recorded at the first Jev hackathon 11 days after launch; TypeSafe DevRel Allie Laabs explains how Jev replaces prompt-parse cycles with typed choice, score and yes/no outputs.

John Lindquist joins Claire Vo to walk through eight Jev use cases, from a real-time voice to-do app to data deduplication with confidence scores, on the episode billed as Jev week at How I AI.

The desk's index of every Jev measurement it has published, keeping each figure's original caveat: JevBench boards, Vercel gateway reach, Celeris and Artificial Analysis results, and the desk's own Jev Day probes which it says it did not rerun.

Side-by-side tables placing similar Jev tests next to each other, from the JevBench API board to Vercel paid-team reach, with the desk's warning that differing sets, labels and comparators mean the rows are not a ranking.

The Register's launch report on TypeSafe AI's Jev: the $40 million raise, typed probabilistic decisions, the Doom demo, the 70-500 ms latency and $0.042/MTok pricing claims, and the structured-output caveat.