Vercel raises Functions package limit to 5 GB on Fluid compute
Vercel raised the maximum package size for Functions on Fluid compute from 250 MB to 5 GB, a 20x increase. The change removes a common deployment blocker for browser automation, larger Python AI stacks, image processing, and heavier backend workloads.

TL;DR
- Vercel raised the maximum Functions package size on Fluid compute from 250 MB to 5 GB, a 20x jump, according to vercel_dev's announcement and rauchg's launch post.
- The rollout is aimed at workloads that previously hit the old ceiling, with vercel_dev's post explicitly calling out Python data and AI libraries, browser automation, image processing, and larger backend apps.
- cramforce's engineer reply framed the 250 MB cap as a recurring deployment pain point, which makes this look less like a vanity number bump and more like a blocker removal.
- The change ships with longer execution windows too: in rauchg's follow-up reply, Vercel said Functions now run for up to 30 minutes on Fluid compute.
- Vercel is also drawing a line on the product shape, with cramforce's clarification saying these bigger Functions are still serverless, not Lambda.
You can open the launch link from Vercel's post and check the Fluid compute link from rauchg's reply. The interesting bit is how broad the target workload list already is: vercel_dev bundled AI Python stacks, browser automation, image processing, and heavier backend apps into the same announcement, while rauchg's reply added a 30 minute runtime ceiling a few hours later.
5 GB packages
The headline change is simple: Functions on Fluid compute can now ship with packages as large as 5 GB, up from 250 MB.
Vercel presented it as a capacity unlock for workloads that were awkward to squeeze into serverless packaging constraints. The official tweet names four categories:
- Python data and AI libraries
- Browser automation
- Image processing
- Larger backend apps
The old 250 MB ceiling
The more useful detail came from cramforce's reply, which says fitting Functions into the old 250 MB limit was "sometimes hard" when deploying to Vercel. That suggests the practical story here is packaging friction, not just raw artifact size.
That framing also explains why the announcement led with dependency-heavy stacks. Python AI environments and browser tooling are exactly the kinds of bundles that get bulky fast.
Runtime and platform shape
The package bump landed alongside a runtime expansion. In rauchg's follow-up, Vercel said Functions now run for up to 30 minutes, and added that "other form factors for long duration are coming."
One other clarification matters: cramforce said the product is still serverless, not Lambda. So the company is widening what fits inside Vercel Functions without repositioning the feature as a different compute product.