Fireworks AI
ai-infra usage-based
AI: bundled
Sources
Insights
Computed from plan data, changelog events, and field coverage.
AI Investment
heavy
Based on AI features, packaging, and credit mechanics
Pricing Complexity
medium
Plan count, add-ons, usage dimensions, credit systems
Transparency
high
Whether prices are publicly visible on pricing page
Pricing Velocity
slow
Number of changelog events in last 12 months
Price Direction
up
Direction of recent price changes from changelog
Field Fill Rate
100%
Percentage of schema fields populated
Changelog 1 event
price increased May 1, 2026
Scheduled price increases: H100 $6->$7/hr, B200 $9->$10/hr
Before: H100 $6/hr, B200 $9/hr
After: H100 $7/hr, B200 $10/hr
Related Companies
Same model (usage-based):
Same category (ai-infra):
Pricing analysis:
Notes
AI inference platform. Serverless by model size: <4B $0.10/MTok, 4-16B $0.20, 16B+ $0.90. Premium models $1.40-$3.48/MTok. On-demand GPU: H100 $6/hr, B200 $9/hr, B300 $11/hr. Cached input at 50% off. Batch at 50% off. Fine-tuned serving at base price (no surcharge). Per-second billing. Price increases scheduled May 2026.
Linked From
Pricing Models