Model developer · Service-channel quotes
Fireworks AI offers a Serverless model API billed on input, output and cached tokens across Standard, Priority and Fast serving paths, plus training fine-tuning and per-GPU-second dedicated deployments. No unconditional free allowance is published.
Usage-based API, monthly subscriptions and sales quotes are billed differently and are not directly compared; the verified ways to buy are listed below.
Price unverified
Serverless model API billed on input, cached input and output tokens from a prepaid balance. Standard, Priority and Fast serving paths price differently and batch inference bills at 50%.
Verified: 2026-11-01 · Official source
Contact sales for pricing
Dedicated deployments and enterprise plans are billed per GPU second or by contract, with higher throughput, lower unit cost and higher rate limits. No unified price is published; contact sales.
Verified: 2026-11-01 · Official source
No verified public personal monthly subscription at this time; the listings above are the verified usage-based API and the custom / enterprise option. Any future subscription will be added after verification.
Usage-based API, monthly subscriptions and sales quotes are billed differently and are not directly compared; the verified ways to buy are listed below.
价格待核实
Serverless model API billed on input, cached input and output tokens from a prepaid balance. Standard, Priority and Fast serving paths price differently and batch inference bills at 50%.
核验: 2026-11-01 · 官方来源
询价
Dedicated deployments and enterprise plans are billed per GPU second or by contract, with higher throughput, lower unit cost and higher rate limits. No unified price is published; contact sales.
核验: 2026-11-01 · 官方来源
No verified public personal monthly subscription at this time; the listings above are the verified usage-based API and the custom / enterprise option. Any future subscription will be added after verification.
Distinct model versions are listed once; aliases are not additional models. Prices below are for API usage.
Platform entry for Fireworks AI. It carries no base-model price of its own; model call prices live on the channel offers.