Kong Reviews
Kong Reviews

Fireworks AI pricing and free tier

Fireworks AI paid plans from $0.008 per 1M tokens (embedding models up to 150M parameters). Read from its pricing page on 25 October 2026.

Categories: LLM APIs, AI inference and GPUs. Website: https://fireworks.ai. Pricing page: https://fireworks.ai/pricing.

ProductPricing modelFree tierLowest paid price
Fireworks AI
Per-model serverless LLM prices are in the pricing docs. On-demand GPUs from $8/hour (H100 or H200); fine-tuning from $0.50 per 1M training tokens (LoRA SFT, models up to 16B).		usage—$0.008 per 1M tokens (embedding models up to 150M parameters)

Caveats

  • Serverless LLM prices per model are in the pricing documentation, not on the pricing page.

Alternatives to Fireworks AI

This page as Markdown

What do I use to build this?
And how much will it cost?What do you want to build?

Describe it in a sentence, like “an app that tracks the beans in my pantry”. You get what to build it with, what it runs on, and what it costs now and as it grows, with a source for every price.Describe it and press Plan. You get what to build it with, what it runs on, and what it costs now and as you grow, after the credits you qualify for.

What Are You Building?

Edit Your Plan

Users

App Type

Budget & Credits

Read a PRD

Prices read from each vendor’s own pricing page

  • Vercel
  • Supabase
  • Stripe
  • Clerk
  • Resend
  • Neon
  • Cloudflare
  • PostHog

317 services listed across 36 categories. 70 priced in full, plan by plan.

One place
for every stack decision.

Describe it. Get one pick per category, and why.

See where the bill jumps before your users do.

Free tiers, plans and limits, side by side.

Already built? Find what to keep and what to change.

What you actually pay, next to what you planned.

See what real products are built on.

Recordings of the actual app. Inputs are illustrative; Billing shows an example account.

Give your AI assistant
access to Kong Reviews.

It plans the stack, prices it at your scale, and hands back package names, environment variables and docs. Your assistant writes the code.

Then keep going in the same chat. Pick up a saved project, price a change before it is built, track setup step by step, and ask why the bill went up.

Free while in preview.

  • Claude
  • ChatGPT
  • Cursor
  • VS Code

I’m building a booking app for yoga studios with sign-in, payments and email reminders. What should it run on, and what will it cost at 20,000 users?

Example prompt. The answer is a live result from recommend_stack, the tool your assistant calls, at today’s prices.
No vendor pays to be listed or ranked higher, and there are no referral links.

Every price change is public: who changed what, when and why.

Already built?
Start from what you run.

Point it at your files, repository or website to see what to keep and what to change. Then bring in what you actually pay, next to what you estimated.

Sign-in required for billing and saved stacks.

Spend less time collecting bills

Import a CSV, upload a text-based PDF, paste an invoice or enter an amount. Recurring subscriptions add completed months automatically.

Supported OpenAI, Anthropic, Vercel, Stripe, AWS and DigitalOcean connections sync completed months nightly. Billing access depends on the provider; personal Claude API keys cannot read organization costs.

Forward receipts when they arrive

When receipt forwarding is enabled, send receipts to your private billing address and review the service, month and amount in Billing. It reads the email body, not attachments; upload PDFs separately.

Revisit decisions when things change

Save a stack or audit and enable alerts for price changes, approaching cost jumps, expiring credits and bills above estimates. Checks run daily using saved usage assumptions and require a confirmed email.

Pick what works.
Build what’s next.

Free while in preview. Create a free account to plan, no card needed.

See pricing

Which vendor should we add?

One vendor per request.

Submit Your Product

Tell us about it here. We check the stack against public sources and take screenshots before anything is listed, and you get a reference to follow it.

Runs on
Add what it's built with (optional)

A public repository lets us confirm the stack sooner.

Prefer email? Email the details to us instead.

Get the most out of Kong Reviews

Filters

Pick what works.
Build what’s next.

Create a free account to plan your stack, save it, and get an email when a price in it changes. No card needed.