Key points
- Vercel data shows Jev reached 35% of paid teams within two weeks of its September 15 limited release, nearly double second-place Sonnet 5.5.
- Jev, built by TypeSafe AI, is a lightweight model that picks from a few predetermined outputs, which makes it cheaper and faster for classification.
- Open source models now account for over 70% of tokens and nearly 25% of spend on Vercel's AI gateway.
- MarkTechPost's headline says TypeSafe AI has released Jev 1.14 with images, a box type and 262k context; no details were available.
Why it matters for marketers: Cheap, narrow classification models are gaining real usage, so teams sorting large volumes of content or data can test them for cost and speed gains over general models.
Vercel published data showing that Jev, a model from TypeSafe AI, was used by 35% of its paid teams within two weeks of launch. That is nearly twice the reach of the next most popular model, Sonnet 5.5, and the fastest two-week adoption Vercel has recorded, according to the company's figures reported by TechCrunch.
Jev was released in limited form on September 15. It is built to choose among a small set of predetermined outputs rather than generate open-ended text. That constraint lets it run cheaper and faster than larger alternatives, which suits sorting or labeling large volumes of data. Vercel itself adopted it early, replacing GPT-6 Luna for classifying harmful commands. A separate headline from MarkTechPost says TypeSafe AI has now released Jev 1.14 with image support, a box type and a 262k context window, though the details were not available to us.
For marketers and content teams, the practical use is in the unglamorous work: tagging assets, routing leads, moderating comments, categorizing support requests or sorting content libraries. Tasks like these rarely need a full generative model, and a cheaper classifier can cut cost and latency in agentic workflows.
One caution: the data comes from Vercel's AI gateway, which is a limited view and may undercount usage of labs like OpenAI and Anthropic that offer their own APIs. The report also shows open source models handling more than 70% of gateway tokens and nearly 25% of spend, while frontier lab spending grew tenfold. Teams should benchmark Jev against their current classification setup before switching, and watch how the 1.14 update and competing offerings develop.
Sources
This brief is our own summary. Read the full reporting at the source.
