Fireworks AI
Fireworks AI provides model inference, reserved throughput and GPU capacity for serving and training.
Is Fireworks AI free?
No free plan is recorded for Fireworks AI. Pricing is usage-based.
Read from Fireworks AI official pricing on .
Pricing
Amounts retain their recorded currency symbols.
| Plan | Recorded price | Billing | Notes |
|---|---|---|---|
| Serverless | $0.05 · usage per 1M input tokens | usage per 1M input tokens | Starting input rate; NVIDIA Nemotron 3.5 Lightning 30B A3B Standard is $0.05 input, $0.01 cached input and $0.20 output per 1M tokens. Model and mode change rates; prepaid credits, no commitment. |
| Reserved Throughput | Custom | custom | Reserved dollars per minute at Serverless rates; commitment required, usage above reservation may bill at standard rates. |
| Fireworks Sovereign | Custom | custom | Enterprise private-preview contract for inference in your cloud account or data center. |
| Embeddings up to 150M parameters | $0.008 · usage per 1M input tokens | usage per 1M input tokens | Not recorded |
| H100 80 GB | $8 · usage per GPU hour | usage per GPU hour | Not recorded |
| H200 141 GB | $8 · usage per GPU hour | usage per GPU hour | Not recorded |
| B200 180 GB | $13 · usage per GPU hour | usage per GPU hour | Not recorded |
| B300 288 GB | $15 · usage per GPU hour | usage per GPU hour | Not recorded |
| GB300 288 GB | $20 · usage per GPU hour | usage per GPU hour | Not recorded |
Alternatives
No recorded competitors are in the catalogue.