Fireworks AI pricing - On-Demand deployments
database
Fields
id
sources/sup-cloud-fireworks-pricingPublisher
Fireworks AI
Kind
database
Accessed
2026-10-03
notes
On-Demand Deployments, 'Pay per GPU second, with no extra charges for start-up times'. H100 80GB $0.134/min = $8.00/h; H200 141GB $0.134/min = $8.00/h; B200 180GB $0.217/min = $13.00/h; B300 288GB $0.250/min = $15.00/h; GB300 $0.334/min = $20.00/h. Note the minute and hour columns are internally inconsistent ($0.134 x 60 = $8.04, printed as $8.00) - the hourly figure is the rounded one. Region-restricted deployments are a stated 1.5x premium. Their serverless per-token pricing is a different unit and is not recorded here. Fireworks is inference-shaped: this rate buys a dedicated deployment, not bare compute.
Notes
On-Demand Deployments, 'Pay per GPU second, with no extra charges for start-up times'. H100 80GB $0.134/min = $8.00/h; H200 141GB $0.134/min = $8.00/h; B200 180GB $0.217/min = $13.00/h; B300 288GB $0.250/min = $15.00/h; GB300 $0.334/min = $20.00/h. Note the minute and hour columns are internally inconsistent ($0.134 x 60 = $8.04, printed as $8.00) - the hourly figure is the rounded one. Region-restricted deployments are a stated 1.5x premium. Their serverless per-token pricing is a different unit and is not recorded here. Fireworks is inference-shaped: this rate buys a dedicated deployment, not bare compute.
Cross-references
This record references no other records.