BREAKING: A ChatGPT co-inventor just came out of 2 years of stealth with a new frontier…
This is a AI post classified by Jev as Frontier models (a model release), kept by the AI Radar because it carries real work, not commentary.
BREAKING: A ChatGPT co-inventor just came out of 2 years of stealth with a new frontier AI model that doesn't generate text. Diogo Almeida built Jev, a new model type from TypeSafe AI, trained with a new method called RLCD. The numbers are sharp: • 20-200x faster than LLMs • 40-400x cheaper (output tokens are free, permanently) • Optimized for decisions and composable intelligence, not chat The trade-off: Jev can't generate text. It's built for code + AI workflows, high-cardinality choices, and decision-making at speed. At ~$42 per billion input tokens with free output, the cost structure
Posted by Guillermo Flor (10.8k followers) 1 h ago · 0 likes · 133 views · view the original post on X. Kept by the AI Radar as Frontier models.
More AI work like this
- Grok 4.7 tokens cost 5x less for input and 8.33x less for output than GPT-6 Astra. — @DanDr1s
- Alibaba introduced exciting new updates to the Qwen model family today at Apsara… — @alibaba_cloud
- 🚨 Qwen are making some good progress on RSI apparently — @LuminaBench
- 第二波实测来了，Mimo-V2.6-Pro(OpenCode) Vs Grok 4.7 (high in Grok build) — @xueyu1125
- omg — @himanshustwts
- 🚨 Qwen4 looks interesting — @LuminaBench
- ALIBABA 🔥: Qwen 4 Max, Qwen 4 Flash, Qwen 4 Plus and Qwen 4 27B models have been… — @testingcatalog
- openai is preparing to release new bot "aeon" — @notjazii
Every post is read and classified by Jev (TypeSafe): what it is, which market it belongs to, and whether the link is a real tool. 41.2k posts from 5k X accounts over the last 14 days, 1.7k tools, 19 markets. Collected every 5 minutes, fully re-ranked every hour — last update 2026-09-22 10:17 UTC. Full method.