Jev ignites a new decision-AI model race — and gets caught in an originality dispute after going viralJev 引爆决策 AI 模型新赛道，爆红后被卷入原创争议

IT Home, Oct. 13 — Tech outlet Tom's Hardware published a post today (Oct. 13) reporting that after TypeSafe released the Jev model it ignited a decision-AI model race; as Microsoft, OpenAI, Cloudflare and Perplexity joined the fray, Jev also faced a plagiarism controversy. According to the post, after TypeSafe released Jev the AI industry quickly produced a wave of similar decision models: Microsoft launched Microsoft Decision, OpenAI launched a Decisions API, Fastino released Glide, while Cloudflare Clef and Perplexity Decider use open weights so users can run them on their own hardware. These models do not generate text token by token and keep no conversation history or long context; they return probability results directly and can process questions in parallel, making them faster and cheaper. TypeSafe previously claimed Jev is 193 times faster than a standard LLM at 1/445th of the cost. Behind the boom, Jev also sparked an originality dispute. ConvAI Innovations founder Nandakishor Mukkunnoth said his SalesRLAgent paper, published March 30, 2025, had already proposed a probabilistic estimation model and openly released model weights and datasets. Mukkunnoth said SalesRLAgent targeted a specific sales scenario, so it never got the same attention as a general decision engine. TypeSafe released Jev as a fully managed product with servers, API, docs, a developer kit and pricing. Mukkunnoth then released an open-source model on the same day Jev launched.