Jev, a decisions-only model, matches gpt-4o-mini 50/50 while being 2.4x faster and 3.4x cheaper
Al_Grigor · x · 2026-09-18
A new model called Jev doesn't generate text — it only makes decisions. You send a state (text or JSON) plus typed questions, and it returns structured JSON with probability distributions.
The author replaced gpt-4o-mini for routing tasks across his 3 agents and benchmarked both on 50 hand-labeled requests: identical 50/50 accuracy, but Jev was 2.4x faster (0.32s vs 0.76s median) and 3.4x cheaper.
It's aimed at anywhere you pay a cheap model for structured-output decisions: routing, guardrails, triage. Full write-up with API and benchmark included; author is running a workshop Tuesday 16:30 Berlin time for AI Shipping Labs members.
Related event: TypeSafe Launches Jev, a Decision-Only Model That Never Writes a Word(44 posts)→
More from coding & agent
- Developer claims running Jev over retrieved chunks fixes RAG precision — NickPassig · 2026-09-18
- Dev replicates local-QA fine-tune idea with RAG using Nigerian data, 3,214 rows cleaned — saheedniyi_02 · 2026-09-18
- Merge Agent Handler Adds Ordinal, Browserbase, Discord and Copper Connectors for AI Agents — shensi · 2026-09-18
- WTF Is Jev? Nine Things People Are Already Building With the Decision-Making AI — blaizedsouza · 2026-09-18
- Qoder Makes Qwen3.8-Flash Free Through Sept 30 and Offers 100 Daily Credits — Scobleizer · 2026-09-18
- Univer ships Office Harness: isolated worktrees let agents parallel-edit connected docs — Scobleizer · 2026-09-18