Runners
GPU GitHub Actions runners
NVIDIA L4 and A10G runners for model evaluation, CUDA compilation and inference tests in CI.
Specifications
- GPUs
- NVIDIA L4 24 GB, A10G 24 GB
- vCPU
- 8 – 32
- Memory
- 32 – 128 GB
- Drivers
- CUDA 12.6, cuDNN 9 preinstalled
Runner labels
runnerhut-gpu-l4runnerhut-gpu-a10g
GitHub does not offer GPU runners. Teams that need them either self-host or skip GPU tests in CI entirely — which usually means the first time a CUDA kernel is exercised is in production.
- Compile CUDA kernels and cache the artefacts between runs
- Run model evaluation suites on pull requests
- Verify inference latency budgets before merge
- Build and test GPU-enabled container images natively
Your next build could be twice as fast, at half the price
Start free. Migrating away is the same one line, and we publish that diff too.