Serverless GPU Benchmarks
Measured runs of a pinned workload per provider — median duration, real-time factor, sample counts, and observed hardware. No synthetic scores.
Results
0 published results · filtered
No benchmark results are published yet.
How to read these numbers
Each row summarizes the currently published campaign for one provider target, scenario, and measurement origin. Values use successful, validated runs: with multiple runs we show the median, while a one-sample row is that single observation. The current workload is a pinned Parler-TTS Mini text-to-speech request. Real-time factor is inference time divided by generated-audio duration; below 1.0× means inference was faster than playback.
- A cold observed scenario is configured to observe scale-from-zero behavior and retains container-session evidence. It is not a provider guarantee that every invocation was cold.
- Time to result is measured by the Cloudflare runner from the beginning of an invocation, before provider submission, until the complete result has been retrieved and validated. Current async profiles poll once per second, so nominal observation lag is up to roughly one second. Older campaigns retain their runner version and may use an earlier profile.
- Median startup is the workload's own application-initialization measurement inside the container. Provider delay can include that initialization, so timing phases may overlap and must not be added together. Provider timings are supporting evidence; time to result remains the user-visible measurement.
- Requests originate from a Cloudflare Durable Object constrained to the EU with a best-effort Western Europe placement hint, not a fixed physical host.
- Compare rows only when the workload version, scenario, and measurement origin match.
- A missing cost is unknown, not free.