Fireworks AI raises $52M Series B to lead industry shift to compound AI systems. Learn more

Skip to main content

Pricing to seamlessly scale from idea to enterprise


Powerful speed and reliability to start your project

600 requests/min rate limit
Up to 100 deployed models
Custom PEFT add-ons
Pay per usage
Free initial credits
Deploy up to 4 GPUs on-demand


Personalized configurations for serving at scale

Everything from the Developer plan
Custom rate limits
Team collaboration features
API telemetry and metrics
Dedicated email support
Deploy up to 16 GPUs on-demand


Personalized configurations for serving at scale

Everything from the Business plan
Custom pricing
Unlimited rate limits
Dedicated and self-hosted deployments
Guranteed uptime SLAs
Unlimited deployed models
Support w/ guaranteed response times

Pricing overview

Fireworks is fully pay-as-you-go, besides enterprise deals. We have multiple pay-as-you-go product offerings, including serverless text model inference, image generation, fine-tuning and on-demand, private GPU inference. Spending on all offerings contribute to spending limits (including credits-based spending) based on past historical usage.

Serverless text models

Base model parameter count$/1M tokens (Applies to both input and output tokens)
0B - 16B$0.20
16.1B - 80B$0.90
MoE 0B - 56B (e.g. Mixtral 8x7B)$0.50
MoE 56.1B - 176B (e.g. DBRX, Mixtral 8x22B)$1.20
Yi Large$3.00

Per-token pricing is applied only for serverless inference. See below for on-demand deployment pricing.

LoRA models deployed to our serverless inference service are charged at the same rate as the underlying base model. There is no additional cost for serving LoRA models.

Image models

Image model, $/stepImage model w/ ControlNet, $/step
$0.00013 ($0.0039 per 30 step image)$0.0002 ($0.0060 per 30 step image)

For image generation models, like SDXL, we charge based on the number of inference steps (denoising iterations). All image generation models besides SD3 are priced identically. SD3 uses pricing from Stability AI.

Multi-modal models

For multi-modal models like LLaVA, each image is billed as 576 prompt tokens.

Whisper (Speech Recognition)

For the Whisper speech recognition model, we charge per second of audio input at a $0.004/minute rate.

Embedding models

Base model parameter count$/1M input tokens
up to 150M$0.008
150M - 350M$0.016

Embedding model pricing is based on the number of input tokens processed by the model.


Model$ / 1M tokens in training
Models up to 16B parameters$0.50
Models 16.1B - 80B$3.00
MoE 0B - 56B (e.g. Mixtral 8x7B)$2.00
MoE 56.1B - 176B (e.g. DBRX, Mixtral 8x22B)$6.00

Fireworks charges based on the total number of tokens in your fine-tuning dataset (dataset size * number of epochs). Fireworks only charges for the costs of tuning - there's no additional cost to deploy fine-tuned models and inference costs are the same as the base model.

On-demand deployments

GPU Type$/hour (billed per second)
A100 80 GB GPU$3.89
H100 80 GB GPU$7.79

On-demand deployments are billed by GPU-second using the above rates. Pricing scales linearly when using multiple GPUs. Users do not pay for start-up times.

Spending limits

Spending limits restrict how much you can spend on the Fireworks platform per calendar month. The spending limit is determined by your total historical Fireworks spend. You can purchase prepaid credits to immediately increase your historical spend. Visit our FAQ for answers to common billing questions.

Note: Credits are counted against your spending limit, so it is possible to hit the spending limit before all of your current credits are depleted.

TierSpending LimitQualification
Tier 1$50 / monthDefault with valid payment method added
Tier 2$500 / month Total historical spend of $100+
Tier 3$5,000 / month Total historical spend of $1,000+
Tier 4$50,000 / month Total historical spend of $10,000+
CustomContact us at [email protected]