Jev, an AI decision model built by TypeSafe AI, has become the fastest-adopted AI model ever recorded on Vercel’s platform. Within two weeks of launch, 35% of paid teams were using it.
Vercel shared the figures on Thursday in its AI Gateway Production Index for October 2026. The company said no model has seen faster adoption in its first two weeks, based on its data.
Jev Pulls Ahead of Sonnet 5.5
According to Vercel, Jev reached nearly double the number of teams as the second most popular model. That model is Sonnet 5.5, made by Anthropic.
TypeSafe AI released Jev in limited form on September 15. Earlier TechCrunch coverage described it as a new kind of model from one of the inventors of ChatGPT.
Jev is built as a lightweight tool. It chooses between a small number of set outputs instead of writing open-ended answers.
Limiting the outputs allows Jev to run faster and at a lower cost than models like Sonnet. This makes it useful for sorting and classifying large amounts of data quickly.
The model drew strong interest from developers soon after its release. OpenAI and Amazon have since launched competing decision models, along with other companies.
Reports of OpenAI’s version appeared on September 30. Amazon released its own version on October 1.
Until this report, there was little hard data showing how widely Jev was being used. Vercel’s figures show that early interest turned into real use.
What Vercel’s Data Covers
The numbers come from Vercel’s AI Gateway. This is a unified API service that lets developers reach many AI models in one place, similar to OpenRouter.
The data offers a limited view of the market. It may undercount large labs such as OpenAI and Anthropic, which sell access through their own APIs.
Vercel was also an early user of Jev itself. The company used it to replace GPT-6 Luna for classifying harmful commands.
The report also tracked open source AI models. These models now make up more than 70% of the tokens processed through Vercel’s gateway.
Open source models also account for nearly 25% of total spending on the gateway. Their share of usage is much larger than their share of spending.
The overall market is growing quickly. Spending on frontier AI labs rose tenfold, according to the report.
Even with that growth, frontier labs now make up a smaller share of total spending on the gateway. Vercel’s data shows spending rising for both frontier and open source models.
