ChatGPT co-inventor launches Jev: claims 20-200x faster, 40-400x cheaper than LLMs
GabGarrett · x · 2026-09-18
ChatGPT co-inventor Milind (@CompleteSkeptic) is launching Jev, a new frontier model trained with a method called RLCD after two years in stealth, positioned as composable intelligence optimized for decisions.
- Claimed specs: 20-200x faster than LLMs, 40-400x cheaper with free output tokens.
- Demo: no screenshots, no LLM, no DOM parsing — a local CoreML model segments every button/UI element on screen, on-device OCR reads labels, and Jev returns probabilities over elements to pick the best click, looping until the goal completes.
- 90ms per decision, faster than any LLM computer-use the tester has tried.
Several developers shared early experiments, calling it a new building block for a wave of apps.
Related event: TypeSafe AI Launches Jev, a Decision-Only Model That Outputs No Text(73 posts)→
More from coding & agent
- One stale date silently broke a WhatsApp reply agent: 19 messages unanswered — petrusenko_max · 2026-11-02
- The 10X engineer is now just a SKILLS.md: 6 workflow rules for AI coding agents — mdancho84 · 2026-11-02
- Auditing my MCP credential broker: the worst hole was the config file — Relevant_Mousse_9117 · 2026-11-02
- AWS's open-source Strands Harness claims 45% cheaper than Claude Code, reigniting harness debate — blaizedsouza · 2026-11-02
- Xiaomi MiMo's CodeSpecWeave trains coding agents on interdependent specs for long-horizon coding — XiaomiMiMo · 2026-11-02
- Sketchy v0.1.6: AI diagrams use 29% fewer output tokens than plain text — Nkgsporty · 2026-11-02