TypeSafe's fast-decision AI Jev sparks rivalry as OpenAI launches Decisions API
TypeSafe AI’s Jev model, built for rapid classification tasks, went viral and prompted OpenAI to release a competing Decisions API within weeks.
Jev, the decision-focused AI from TypeSafe AI, gained rapid traction among developers after its September launch, with Vercel reporting that almost 13% of its paid teams used it in the first 24 hours, double the rate of any prior model debut. The system is designed to classify data—such as sorting customers or flagging email errors—more cheaply and faster than general-purpose large language models, while still allowing natural-language instruction.
Three weeks later, OpenAI introduced a Decisions API built on GPT-6 Luna, promising up to ten-fold faster responses and image input capability, though it costs $0.10 per million tokens versus Jev’s $0.042. TypeSafe’s CEO Diogo Almeida, a former OpenAI engineer, says Jev fills a gap between answering questions and automating routine code-driven tasks. Early adopters like Niels Mouthaan and Ashutosh Mathore praise Jev’s speed for spelling checks and compliance software, while others note the need to craft prompts differently from traditional LLMs. OpenAI’s entry leverages its large user base and ecosystem, but TypeSafe’s two-year development and focus on reliability may give it an edge in production settings.
Why it matters
The race to cheaper, faster AI decision tools could reshape how businesses automate routine tasks.
In this story
