Ex-ChatGPT co-inventor launches Jev, claims 20-200x faster and 40-400x cheaper
pranavmarla · x · 2026-09-17
A self-described ChatGPT co-inventor launched Jev, a frontier model trained with a new method called RLCD, claiming 20-200x faster inference, 40-400x lower cost (free output tokens), and decision-optimized composable intelligence. A Stanford CS PhD benchmarked Jev/Typesafe on HiringCafe (2.5M MAU consumer app) scoring resume-to-job-description relevance, publishing a detailed thread.
Related event: New Jev model claims 20-200x speedup; replication on Qwen 4B raises doubts(2 posts)→
More from Models
- "As smart as terra": the biggest fumble of the jev rollout — willcb · 2026-09-17
- Daily AI Brief: Claude merges Chat and Cowork, OpenAI ships misalignment disclosure framework — testingcatalog · 2026-09-17
- NVIDIA's SpatialClaw uses code as action interface, beats prior agent by 11.2 points on 20 benchmarks — CMHungSteven · 2026-09-17
- 15 Small Models That Beat Models 100x Their Size at One Task — bigaiguy · 2026-09-17
- Early Looks: DeepSeek V4.1 Crushes Agentic/Coding Tasks but Regresses on Some Benchmarks — teortaxesTex · 2026-09-17
- Dev generates a full software promo video in pure code with GPT-6 Astra, no generative models — op7418 · 2026-09-17