Vercel Functions can now run up to 30 minutes
30 minutes. Vercel just unlocked long-running AI workloads—LLM reasoning, document processing, web scraping—without the infrastructure headache.
Why it matters
Vercel's extended function duration removes a critical friction point for AI-native applications. Developers can now run complex LLM reasoning, multi-step agent workflows, and document processing natively on serverless, with billing that pauses during I/O waits—directly enabling AI product builders to ship without custom infrastructure.
The key facts
7 to knowFunction execution duration extended from 800 seconds (~13 minutes) to 30 minutes for Pro/Enterprise
2.1x increase in max execution time
Node.js and Python runtimes supported; additional runtimes coming
Fluid Compute: active CPU billing pauses during I/O (API calls, database queries, model inference)
Use cases explicitly listed: LLM reasoning, streaming AI responses, document/media processing, OCR, web scraping, workflow handlers
Durations >800s currently in beta and require Fluid Compute opt-in
Published June 15, 2026
Go to the source
Vercel Blogvercel.com
Publisher excerpt: Vercel Functions using the Node.js and Python runtimes now support execution durations up to 30 minutes for Pro and Enterprise teams, more than 2x the previous 800 second limit. Support for additional runtimes is coming soon. Use longer-running Functions for work that needs more time to finish,…