> Fireworks AI
Fireworks AI is a self-serve inference platform: per-token serverless (Standard/Priority/Fast tiers) with $1 in free credits and postpaid billing, fine-tuning (LoRA and full-param, priced per 1M training tokens), a shared-pool Serverless Training API, and on-demand GPUs from $8/hr (H100, from Sep 1 2026) to $20/hr (GB300).
// pros
- Self-serve serverless with $1 in free credits, postpaid billing and no cold starts — the on-ramp is frictionless.
- One account spans the full stack: serverless inference, managed fine-tuning, serverless training and on-demand GPUs (H100 $8/hr, B200 $13/hr, GB300 $20/hr from Sep 1).
- Fine-tuned models are served at the same price as their base model — no premium for your own weights.
- Serverless Training API runs LoRA on a shared, always-on trainer pool billed per prefill/sample/train token — no provisioning, no idle cost.
- Published embeddings pricing as low as $0.008/M input on small models.
// cons
- $1 in free credits is onboarding money, not a working tier — budget for postpaid billing from the first serious request.
- The same model costs differently across Standard/Priority/Fast serverless tiers — any price comparison must name the tier.
- GPU prices stepped up on Sep 1, 2026 (H100 $7→$8, B200 $10→$13, GB300 $18→$20 per hour) — re-quote before committing long dedicated runs.
- Region-restricted deployments carry a 1.5x premium.
A strong default for open-weight serverless work that may later need training: prototype on the $1 credit, stay on serverless until throughput demands it, then use the same account's on-demand prices as your anchor.
written by Mike Fleming, Uprouter editorial · verified against official sources September 16, 2026
Value taken from the provider’s own published free-tier terms, expressed in USD.
note: Official pricing page: 'Get started with $1 in free credits' (confirmed 2026-09-16).
| Plan | Type | Monthly | Input /1M | Output /1M | Note |
|---|---|---|---|---|---|
| Free tier (source-reported) | free | — | — | — | New users get about $1 in free credit (roughly two months for small-scale use); without a payment method the limit is 10 req/min, rising to 6,000 req/min after adding a card; pay-as-you-go once credit is used up. |
| Free tier | free | — | Free | Free | $1 one-time |
| Pay-as-you-go | payg | — | $0.900 | $0.900 | llama-3.3-70b class; per-model rates |
| Subscription | subscription | — | — | — | serverless pass; check current terms |
| Llama-3.1-405B | payg | — | $3.00 | $3.00 | Serverless inference rate. |
| DeepSeek-V3 | payg | — | $0.900 | $0.900 | Standard serverless tier. |
Pricing normalized from public sources — always verify with the provider.
// model prices
| Model | Input /1M | Output /1M | Context |
|---|---|---|---|
| Llama family | — | — | — |
| DeepSeek family | — | — | — |
| Qwen family | — | — | — |
$ Does Fireworks AI have a free tier?
Yes — Fireworks AI offers a free tier that we value at roughly $1.00 in free credits. Always confirm current limits on Fireworks AI's official pricing page — free tiers change without notice.
$ How much does Fireworks AI cost per 1M tokens?
Fireworks AI does not publish simple per-1M-token pricing (it may bill per request, per output, or via a subscription plan). Check their official pricing page for exact figures.
$ Is Fireworks AI safe to use?
Uprouter rates Fireworks AI as low risk. Low risk: a first-party platform with one published pricing page that states the free credit, tier structure, training prices and a dated GPU price change explicitly. Nothing hidden; the cost risks are tier selection and the Sep 1, 2026 on-demand price step-up. Risk ratings are editorial, evidence-based, and never influenced by affiliate relationships.
Raw community sentiment, shown unblended. Votes never feed into objective data fields like price, risk, or status — and never into sort order.
No published reviews yet — be the first.
// more_informationprice history · live status · risk & confidence · code snippets
price_history
No recorded changes yet.
live_status
full history →best-effort probes · not a guarantee · last 15h ago
| Checked | Status | Latency |
|---|---|---|
| 15h ago | up | 166ms |
| 2d ago | up | 199ms |
| 3d ago | up | 215ms |
| 4d ago | up | 149ms |
| 6d ago | up | 284ms |
| 7d ago | degraded | 335ms |
| 9d ago | up | 168ms |
| 11d ago | up | 242ms |
| 12d ago | up | 218ms |
| 13d ago | up | 292ms |
| 14d ago | up | 178ms |
| 15d ago | up | 174ms |
| 16d ago | up | 186ms |
| 16d ago | up | 320ms |
| 17d ago | up | 171ms |
| 18d ago | up | 227ms |
| 19d ago | up | 132ms |
| 19d ago | up | 168ms |
| 20d ago | up | 202ms |
| 21d ago | up | 210ms |
| 22d ago | up | 250ms |
| 22d ago | up | 337ms |
| 23d ago | up | 173ms |
| 24d ago | up | 214ms |
| 25d ago | up | 207ms |
| 26d ago | up | 189ms |
| 27d ago | up | 207ms |
| 27d ago | up | 164ms |
| 28d ago | up | 264ms |
| 29d ago | up | 180ms |
Uprouter runs its own probes; best-effort, not a guarantee. Probe results depend on our network location and cadence — for production, always use the provider’s own status page.
risk_&_confidence
Low risk: a first-party platform with one published pricing page that states the free credit, tier structure, training prices and a dated GPU price change explicitly. Nothing hidden; the cost risks are tier selection and the Sep 1, 2026 on-demand price step-up.
Confidence reflects how much verifiable evidence (official docs, probe history, community reports) backs this entry. Risk is our editorial assessment of operator reliability and terms — not financial advice.
code_snippets
# Fireworks AI's own API — get a key from their dashboard
curl https://YOUR-PROVIDER-ENDPOINT/chat/completions \
-H "Authorization: Bearer THEIR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "llama-family",
"messages": [{ "role": "user", "content": "Hello" }]
}'Direct to the provider — Uprouter receives nothing; verify their current pricing before purchasing.
Ranked by neutral similarity signals only — live status, free-tier shape and free-credit proximity. Commercial relationships never influence this list.