Puzl's pipeline jobs beat ordinary cloud instances due to burstable compute resources: only pure resource consumption is paid.
* Comparison based on
the benchmarks below against
GitLab.com SaaS runners.
Do not pay for the resource limits during idle time.
Puzl counts cpu-seconds and memory GB-seconds only when your job is actually utilising them.
Compare features | Free | Business | Enterprise |
Dynamic load-based billing: pay for CPU-seconds and memory-seconds actually utilized by your process | |||
Each job is running in its own clean, isolated virtual machine | |||
Maximum concurrent jobs | 10 | unlimited | unlimited |
vCPU allocated for each job | 12 | 48 | custom |
Memory allocated for each job (GB) | 32 | 96 | custom |
Flexible Persistent Data Storage included (GB). Use it to share any data among your jobs. | 0 | 10 | custom |
Job cache, ready to use with your job provider (GitLab, GitHub, etc.) | |||
Use GPU in your jobs | |||
Integrations (use them to isolate different teams, projects or organizations) | 1 | 3 | custom |
Job runners per integration | 1 | 3 | custom |
Manage your runners via dashboard or declarative API | |||
Ticket-based support | 1 business day | Priority support | |
Dedicated nodes for extra isolation of your workloads | |||
Choose the region to run your jobs | |||
GitLab CI | |||
Maximum service containers per job | 2 | 10 | custom |
Ephemeral storage allocated for each job (GB) | 100 | 100 | custom |
Interactive Web Terminal, allowing real-time debugging of pipelines in your GitLab dashboard | |||
GitHub Actions | |||
Ephemeral storage allocated for each job (GB) | 150 | 150 | custom |
GitHub environment fully compatible with the official GitHub Actions runners | |||
Coming soon | |||
Local Docker layer cache for even faster jobs | |||
MS Windows execution environment | |||
ARM-based compute nodes |
© GPU Computing OÜ, 2025
Harju maakond, Tallinn, Kesklinna linnaosa, Rotermanni tn 6, 10111 Estonia
Company code 14476114
VAT number EE102212851