Vercel Frontend cloud platform behind Next.js, v0, and the AI SDK.
Vercel runs its v0 composite code-generation models and its Auto Fix model on Fireworks AI, using speculative decoding to speed up token generation. Shipping support for newly released models within hours rests on an eval foundation in Braintrust, which the team uses to benchmark each new model against the ones already in place and to read production traces for problems customers never report directly. Vercel and Braintrust also collaborate on representing richer AI API features — tool calling, memory, planning — on both the client and the observability side.
vercel.comSources last verified Nov 9, 2026
Fireworks AI Runs v0 composite and Auto Fix models with speculative decoding.
Receipts · 1
“Both Vercel’s Auto Fix model and its v0 composite model uses Fireworks’ Speculative Decoding to speed up token generation.” — fireworks.ai, Nov 3, 2025
Source last verified Jul 27, 2026
Braintrust Benchmarks new model releases and reviews production traces for quality issues.
Receipts · 1
“Braintrust gives the team a structured way to benchmark new models against existing ones” — braintrust.dev, Nov 2026
Source last verified Nov 9, 2026