TypeSafe AI · モデルファミリー

Jev

TypeSafe AI decision model for typed choices, scores and probabilities. No generated prose; input-token billing applies. Evaluate accuracy on your own workload.

APIのみ テキスト 現在:Jev 1.13

現在のリリース

Jev 1.13

安定版

Current documented version checked September 24, 2026. Outputs typed decisions, not prose. Official announcement dates the Jev family launch, not a separately verified release date for this version.

リリース日日付未提供
コンテキストウィンドウ64Kトークン
ライセンスProprietary hosted API
ウェイト未公開

検証済み仕様

Input price USD 0.042 per million input tokens 出典 →
Output billing Output tokens are free; input is billed 出典 →
State plus longest question 32k tokens; total request budget 64k 出典 →
Output types Choice, Score, Noul; typed decisions, not generated text 出典 →
Language support English is primary; test non-English workloads independently 出典 →
Parameter count Not disclosed in reviewed documentation 出典 →

リリース年表

Jev 1.13
安定版 · 日付未提供

Current documented version checked September 24, 2026. Outputs typed decisions, not prose. Official announcement dates the Jev family launch, not a separately verified release date for this version.

コンテキスト 64K APIのみ

公式発表 →

プロジェクト

Official agent skills for designing applications around TypeSafe typed decisions. Development guidance complements the runtime SDKs and API.

★ 2K · ⑂ 109

Integrate TypeSafe decisions into JavaScript and TypeScript services. Answer types are inferred from questions; the package supports ESM and CommonJS.

★ 231 · ⑂ 25

Connect Python services to the TypeSafe API using typed questions and results. The repository provides the client library rather than model weights.

★ 218 · ⑂ 30

関連記事

Jev是什么:Almeida团队推出只做结构化判断的AI模型 9月 24, 2026 把Jev接入应用:官方Python、JavaScript SDK与Agent工具链 9月 24, 2026 Jev适合什么、不适合什么:中文、长上下文与判断阈值 9月 24, 2026