Decision Model Jev Goes Free on Venice API: Typed Answers, No Prose, No JSON Wrangling
0xAllen_ · x · 2026-09-20
Anonymous inference platform Venice announced that Jev, a "decision model" by typesafeai, is now free on the Venice API (beta) via a dedicated /api/v1/decisions endpoint.
Jev inverts the usual LLM pitch: "it answers, it doesn't write." Apps send a shared state (up to 64K tokens) plus questions with predefined answer types—yes/no, choice, or rubric scoring—and the model evaluates each question in parallel and in isolation against the same state, returning machine-ready probabilities, confidence, options, or scores. No coaxing JSON out of a chat model, no schema repair. Answer types are labeled Noul · Choice · Score.
On the Venice side, one key reaches chat, image, video, audio, and Jev, with no state retention, no prompt retention, and no training on your data. The beta's availability and request/response schemas may still change.
More from coding & agent
- 58 law firms bought AI, only one can tie spend to a matter — jkubicki · 2026-09-20
- Dev's LLM-built bot climbs to #1 on ThursdayArena, sharing the prompt template — tetsuoai · 2026-09-20
- A $36,000 Cloudflare Bill Sparks Warning: No Hard Spend Limit, Set Billing Alerts — ThePeterMick · 2026-09-20
- How I run my AI agents on a VPS with Tailscale, key-only SSH and per-agent isolation — EXM7777 · 2026-09-20
- Qwen 27B one-shot prompt builds three playable Super Mario clones — EcstaticDentist · 2026-09-20
- Jev Classifies 1.6k Bookmarks in 22s, 155x Faster and 10x Cheaper Than GLM 4.7 Flash — iannuttall · 2026-09-20