Pricing
Four products. Independent commercial paths.
Start with the product you need. Forge, Compute, Token Optimisation and CI Autopilot can be adopted independently.
Product 01
Token Optimisation
Prepaid processed-input-token packages. See the LLM cost optimization guide for the measurement framework.
Product 02
Varion Forge
Free trial plus monthly higher-capacity plans.
Loading Forge plans…
Product 03
Varion Compute
10 free GPU-hours, then simple usage-based pricing that gets cheaper as you scale.
Free trial
Start 10-hour free trial10 free GPU-hours
Test Varion Compute on your own workload with full metered access. No credit card required.
Up to 1,000 hours
$0.80/GPU-hour
For startups and smaller production workloads.
Start free1,001–5,000 hours
$0.70/GPU-hour
Automatic volume pricing as monthly usage grows.
Start free5,001–20,000 hours
$0.60/GPU-hour
Lower unit pricing for larger production workloads.
Start free20,000+ hours
Custom
Commercial pilot and enterprise rollout terms. Target pricing around $0.50/GPU-hour, based on rollout and volume.
Talk to salesProduct 04
Varion CI Autopilot
Free analysis first. Commercial terms tied to verified realized savings.
Savings-first
Challenge VarionProve the saving before you buy
Run the public CI analysis free. Enterprise deployment can use success-based commercial terms, with the exact fee agreed against verified savings.