Lower pricing with Active CPU pricing for Fluid compute
By Steven Van ·
Fluid compute now bills Active CPU time separately from memory, cutting example costs from $0.318 to about $0.149 per hour.
Vercel Functions on Vercel's Fluid compute now use Active CPU pricing, charging for CPU only while a function is actively computing rather than for the whole time it stays in memory. That removes cost during idle periods, which matters for workloads like LLM inference or long-running AI agents that spend much of their time waiting on external responses.
The model is built on three metrics: Active CPU time (priced at $0.128 per hour), Provisioned Memory ($0.0106 per GB-hour), and per-call invocations. As an example, a Standard machine size running at 100% active CPU now costs about $0.149 per hour, versus $0.31842 per hour under the previous model. Active CPU pricing is enabled by default for Hobby, Pro, and new Enterprise teams; existing Enterprise customers' availability depends on their current plan. The change takes effect after a redeploy.
Read the full announcement for more on how Fluid compute's Active CPU pricing works.
