What would your infrastructure cost on Gameye?
Model your load profile and compare Gameye against Edgegap and AWS — egress included on Gameye.
How many more AWS vCPU-hours your game needs to match Gameye. Default 1.15× is the clock ratio: Gameye's 4.7 GHz fleet average ÷ c8a's 4.1 GHz. It's an assumption that depends on your game, so benchmark to confirm.
| Provider | Compute | Egress | Total |
|---|---|---|---|
| Gameye4.7 GHz avg bare metal · egress incl. | — | Included | — |
| Edgegap$0.069/vCPU-hr · $0.0051/GB-hr RAM · $0.10/GB egress | — | — | — |
| AWS c8a.xlargeSelf-managed EC2 (Pragma/Nakama/custom) · AMD EPYC 9R14 @ 4.1 GHz ★ closest match | — | — | — |
| AWS GameLift (managed)c8a.xlarge at $0.2655/hr on demand · gen-6+ bandwidth included · no reserved pricing | — | Included | — |
—
—
AWS GameLift vs self-hosting on EC2. AWS made outbound bandwidth free on AWS GameLift (gen-6+) instances in June 2026, which the AWS GameLift row reflects. The AWS c8a.xlarge row models self-managed EC2 — the common setup when Pragma, Nakama, or a custom backend allocates your own dedicated servers — where standard EC2 egress (~$0.09/GB) still applies and still scales with players. Gameye includes bandwidth either way.
Edgegap charges roughly $0.069/vCPU-hr for compute, a per-GB-hour RAM charge, and $0.10/GB egress on top — so bandwidth is billed separately and grows with your player traffic. Gameye's $0.07/vCPU-hr includes all egress.
Both AWS rows reflect two adjustments. First, a performance factor you can set with the slider. The default (1.15×) is the clock ratio between Gameye's 4.7 GHz fleet average and c8a's 4.1 GHz. Real throughput depends on your game, so treat it as an assumption and benchmark to confirm. Second, instance packing density: the c8a.xlarge has fixed specs (4 vCPU / 8 GB) — when your container's RAM demand constrains how many fit per instance, idle vCPUs are wasted and the effective cost per container rises accordingly.
Billed per vCPU-hour. Rate depends on container RAM allocation.
avgCcu = peakCcu × avgCcuPct
vCpuHrs/mo = (avgCcu ÷ playersPerContainer) × containerVcpu × 24 × 30
compute = (reservedHrs × $0.035) + (onDemandHrs × $0.07–0.10)
egress = included
total = compute Billed per vCPU-hour plus a separate per-GB-hour RAM charge. Egress billed at $0.10/GB.
compute = vCpuHrs/mo × $0.069
ram = containerHrs/mo × containerRamGb × $0.0051
egress = egressGb/mo × $0.10
total = compute + ram + egress Instance-based billing. Packing density determines how many containers fit per instance. A performance factor (slider, default 1.15× = 4.7 GHz ÷ 4.1 GHz) scales the vCPU-hours AWS needs. It's an assumption: benchmark your own build to confirm.
packing = min(floor(4 ÷ containerVcpu), floor(8GB ÷ containerRam))
effectiveRate = $0.17168/hr ÷ (packing × containerVcpu)
adjustedHrs = vCpuHrs/mo × perfFactor
egress = first 10 TB at $0.09/GB, beyond at $0.085/GB
total = (adjustedHrs × effectiveRate) + egress Same c8a.xlarge instance and packing as the EC2 row, at GameLift's on-demand rate from the AWS price list. GameLift has no reserved pricing, so reserved hours are also billed on demand. Bandwidth is free on gen-6+ instances.
effectiveRate = $0.2655/hr ÷ (packing × containerVcpu)
adjustedHrs = vCpuHrs/mo × perfFactor
egress = included
total = adjustedHrs × effectiveRate Average CCU methodology follows the AWS GameLift calculator convention: average CCU ≈ 30% of peak CCU across a full month, accounting for daily rise and fall in player counts. Adjust the slider to model your specific utilisation profile.
How much does Edgegap cost?
Edgegap bills three line items: about $0.069 per vCPU-hour of compute, roughly $0.0051 per GB-hour of RAM, and $0.10 per GB of egress (bandwidth) on top. There's no flat sticker price — what you pay tracks compute hours, RAM, and data transfer. Because egress is metered separately, the bill grows with your player traffic; at scale, bandwidth alone is commonly 30–50% of the total.
| Cost component | Edgegap | Gameye |
|---|---|---|
| Compute | ~$0.069 / vCPU-hr | $0.07 / vCPU-hr ($0.035 reserved) |
| RAM | ~$0.0051 / GB-hr (billed separately) | Included in the vCPU rate |
| Egress / bandwidth | ~$0.10 / GB | $0 — included |
| Pricing model | Usage-based (compute + RAM + egress) | Per vCPU/hr, egress included |
The headline compute rates are close — the difference is the egress line. On Edgegap, bandwidth is metered at ~$0.10/GB and scales with your success; on Gameye it's zero. Model your own player counts in the calculator above, or see the full Gameye vs Edgegap comparison.
Frequently asked questions
How much does Gameye cost vs AWS GameLift?
Gameye charges $0.07/vCPU/hr ($0.035 reserved) with no egress fees, on every provider and region. AWS GameLift includes bandwidth free on its own gen-6+ instances since 2026; if you self-host servers on EC2 with a backend like Pragma or Nakama, standard EC2 egress (~$0.09/GB) still applies. The bigger lever is per-core performance — Gameye's 5 GHz bare metal hosts more players per core than virtualised EC2 vCPUs. Use the calculator above to model your specific load, or see the AWS GameLift pricing breakdown for instance-by-instance rates.
What's included in Gameye pricing?
Everything: compute, bandwidth, DDoS protection, the Admin Panel, log streaming, warm pools, matchmaker integrations, and the full orchestration API. No egress fees, no per-session charges, no bandwidth surcharges.
How are egress fees calculated on other platforms?
AWS charges ~$0.09/GB for the first 10TB on standard EC2 data transfer. Note the distinction: AWS made bandwidth free on AWS GameLift (gen-6+) instances in 2026, but if you self-host dedicated servers on EC2 (the common Pragma, Nakama, or custom-backend setup) standard EC2 egress still applies and scales with players. Gameye includes all data transfer at no extra charge, on every provider and region.
What is the difference between on-demand and reserved pricing?
On-demand ($0.07/vCPU/hr) is pay-as-you-go with no commitment. Reserved ($0.035/vCPU/hr) is a committed capacity block at a lower rate. Most studios use a mix: reserved for baseline player load, on-demand for peaks and launches.
Does Gameye charge per session?
No. Gameye bills per second of active compute. There are no per-session fees, no minimum session lengths, and no connection charges.