Jev directory
CommunityJev Integrations & Adapters
Jev 1.13 on Opper
Opper's model page for Jev 1.13: one route through the Opper gateway at $0.04 per million input tokens, text-only input with a 64K window, RLCD-trained calibrated probabilities, and a Pac-Man benchmark run across 100 games.
- Reading time
- 3 min
- Published by
- Community
- Author
- Opper
- Last verified
- Oct 2026
Used inJev for Software & DevTools
- Platforms
- API
- Capabilities
- Requires API key
Highlights
- Model id typesafe/jev-1.13.0 on one Opper route, at $0.04 per million input tokens and $0 per million output tokens.
- Input is text only, with 64K tokens per request and at most 32K for the state plus the longest question; English is strongest.
- Questions are evaluated in parallel and in isolation against the same state, so adding questions barely changes latency.
- In Opper's JevMan run, Jev 1.13 played 100 Pac-Man games, ranked equal first of eight, mean score 2,750, best game 6,380.
- Zero data retention is not established for the TypeSafe route; no training on customer data and a GDPR DPA available, verified 18 September 2026.
Watch out
Opper's own listing and own benchmark runs, updated 10 October 2026: the Pac-Man scores are a game benchmark rather than task accuracy, zero data retention is marked not established, and the free tier covers only free models, with a card needed for Jev. Prices may change.
Questions about this entry
Short answers drawn from the linked project or page.
POST to https://api.opper.ai/v3/compat/v1/systemone with your Opper API key, model typesafe/jev-1.13.0, a state, and typed questions; a free account gets an API key straight away, and a card unlocks premium models.
The listing shows $0.04 per million input tokens and $0 per million output tokens on its single route, so you effectively pay for the context it evaluates rather than generated text. Jev is a premium model on the gateway.
64K tokens per request in total, with at most 32K reserved for the state plus the longest question. Input is text only, as a string, JSON object, or array, and accuracy is strongest in English.
Opper's Pac-Man benchmark for decision models: Jev 1.13 played 100 games against four arcade ghosts, ranked equal first of eight, and posted a mean score of 2,750 with a best game of 6,380, level with GPT-6 Luna Decisions and Clef-flash within the margin of error.
Zero data retention is not established for the TypeSafe route; the listing records no training on customer data, a GDPR DPA available, and an unknown transfer mechanism, verified by Opper on 18 September 2026 in the United States region.
Noul returns the probability that a statement is true, Choice picks from options you define, and Score places the state on your rubric; the page suggests ticket classification and routing, intent detection, guardrails, and rubric scoring instead of an LLM judge.
The scores come from Opper's own runs and measure a game, not decision accuracy on real tasks; the page lists no task evaluations, so production thresholds still need testing on your own representative data.
Reactions & coverage
Posts, threads, and videos about this entry from around the web.
X: Jev 1.14 is now on OpenRouter
X: Jev lands on the Vercel AI Gateway
X: Jev as an LLM guardrail inside Agentgateway
YouTube: Greg Isenberg and Ryan Vogel on the businesses Jev unlocks, and a path in through the Vercel AI Gateway.
More Jev resources like this
Pydantic's AI Gateway forwards Jev natively as a bring-your-own-key TypeSafe provider, so one gateway key covers spend views, limits, and input guardrails while TypeSafeModel turns each output field into a question.
Jev Integrations & AdaptersArticle#community#pydantic#gateway
Community
LiteLLM serves decision models on two routes that share one provider list: /v1/decisions for OpenAI-format bodies and /v1/systemone for System One bodies, with Jev available as the typesafe/jev-latest provider.
Jev Integrations & AdaptersDocs#community#litellm#gateway
Community
Concentrate's beta /v1/decisions endpoint serves TypeSafe's Jev with the same request and response shapes as TypeSafe's System One API and OpenRouter, returning calibrated probabilities for noul, choice, and score questions.
Jev Integrations & AdaptersDocs#community#gateway#api
Community
OpenCode's Zen gateway serves Jev 1.13 on its own /v1/systemone endpoint with an OpenCode key: jev-1.13 paid and jev-1.13-free for a limited time, documented with noul, choice, and score examples.
Jev Integrations & AdaptersDocs#community#opencode#gateway
Community
A decision-model marketplace and playground: one POST /v1/systemone endpoint in front of two dozen models including Jev 1.13 and 1.14, with structured answers, probabilities, and a 100-credit starting balance.
Jev Integrations & Adapters#community#marketplace#gateway
Community
LiteLLM's proxy adds TypeSafe guardrail hooks and a Jev complexity classifier for policy judgments and request routing across 100+ model providers.
Jev Integrations & Adapters#community#python#gateway
Community
BerriAI/litellmPython
BerriAI/litellmPythonFrom the Jev guides
Original write-ups on what this entry is for.
From the Jev AI model community
Posts from builders shipping with Jev right now.