Dev Forces Judge Model Jev to Talk at 3 tok/s

Developer Rogerio Chaves released CaveJev, fitting Typesafe's judge model Jev with a caveman vocabulary to make it speak like an LLM. The result: only about 3 tokens per second and higher costs, proving it's a poor chat model.

2026-09-18 ~ 2026-09-18 · 3 related posts

• CaveJev: forcing the Jev judgment model to speak word by word like an LLM — _rchaves_ · 2026-09-18
• Developer turns judgment model Jev into a chatbot: 3 tok/s and pricier than LLMs — _rchaves_ · 2026-09-18
• Chatting With Jev as a Regular LLM: A 3 tok/s Caveman-Style Experience — _rchaves_ · 2026-09-18