NVIDIA GPUs for open model inference
RunInfra publishes rate and capacity facts for this GPU. No measured optimization package is published for this GPU yet. Every unavailable field stays unpublished rather than estimated.
View Model APIsGPU time is not currently sold. Historical rates are reference values, not an offer. Active-tier rates refer to reserved deployments. Rate source: deploy catalog, captured Aug 9, 2026.
Browse the benchmarks for GPUs that carry measured evidence, or use the cost calculator to compare cloud rental prices for A100 40GB.
A measured page ties one published package to its recorded engine and request conditions.
It shows baseline and optimized observations from the same protocol.
It also carries the package quality evidence and verification date.
GPU time is not currently sold. These are historical reference values, not an offer. Active-tier rates refer to reserved deployments. RunInfra publishes an on-demand deploy rate of $2.10/hr for this GPU. Active hourly rate not published. Rate basis: the on-demand deploy rate catalog. This page was derived from that catalog on August 9, 2026.
No. RunInfra has not published a measured optimization package for this GPU yet. Measured results are on the benchmarks page and measured GPU pages.
Use a workspace API key and pay for input, cached input, and output tokens.
View Model APIs