Fireworks AI pricing in 2026

Inference and fine-tuning platform for open-source generative AI models. How Fireworks AI charges, its plans and the alternatives.

Track Fireworks AI and every other subscription in Surba · 14-day free trial Check your stack's total cost

Billing

Pay as you use

Plans

How Fireworks AI prices

Fireworks AI is usage-based: you pay per transaction, message or unit used rather than a fixed subscription, so there's no plan price to show. Surba tracks what you actually pay each month.

Prepaid credits, pay-as-you-go. Serverless per-token rates are in docs (not on pricing page). On-demand GPUs: H100 $8/hr, B200 $13/hr. Embeddings from $0.008 per 1M tokens; LoRA SFT from $0.50 per 1M. Enterprise: contact sales.

Checked on Fireworks AI's site on 26 Oct 2026.

Alternatives

Alternatives to Fireworks AI

Similar tools in Surba's catalog, cheapest first.

All 32 Fireworks AI alternatives ranked by price

FAQ

Fireworks AI pricing questions

How does Fireworks AI charge?

Fireworks AI is usage-based: you pay for what you use (transactions, messages, compute and so on) rather than a fixed subscription, so the bill depends on volume. Prepaid credits, pay-as-you-go. Serverless per-token rates are in docs (not on pricing page). On-demand GPUs: H100 $8/hr, B200 $13/hr. Embeddings from $0.008 per 1M tokens; LoRA SFT from $0.50 per 1M. Enterprise: contact sales.

More AI tools

Every AI tool

Track Fireworks AI and every other subscription in one place.

Surba records what you pay Fireworks AI each month next to your subscriptions, shows your whole software spend in your currency, and flags overlapping tools.

Start your 14-day free trial Check your stack's total cost