← All providers
Fireworks AI pricing
verified 2026-09-02Fast serverless inference for the popular open-weight models, with a production focus and per-token pricing.
data residency & retention
from Fireworks AI's own policies · verified 2026-09-02Headquarters
🇺🇸 United States (US)
Where inference runs
US, EU, Other
Default serverless is global Fireworks-managed infrastructure with no per-request geography disclosed; dedicated deployments publish ~26 named regions (US, EU Frankfurt/Iceland, Canada, APAC).
Region pinning
Yes, self-serve
Self-serve US-only serverless endpoint (us.api.fireworks.ai, at a price premium) with US-resident models; EU-only serverless is enterprise/contact-sales; dedicated deployments can be pinned by region.
Prompt retention
Not retained
Zero Data Retention by default — prompt and generation data is not logged or stored for open models without explicit opt-in; prompt-cache data lives in volatile memory for minutes; Responses API conversation data auto-deletes after 30 days.
Trains on API data
No
Terms commit: customer content is not used to train Fireworks' own models or to improve the service.
Zero data retention
Available
ZDR is the contractual default for inference (carve-outs: safety screening, Responses API, and training/fine-tuning features).
current models · $ per 1M tokens
| Model | Input /1M | Cached in | Output /1M | Context |
|---|---|---|---|---|
OpenAI GPT OSS 120B GPT-OSS | $0.15 | $0.01 | $0.60 | — |
GLM 5.3 GLM | $1.40 | $0.26 | $4.40 | — |
GLM 5.3 Flash GLM | $0.15 | $0.03 | $0.50 | — |
GLM 5.2 GLM | $1.40 | $0.14 | $4.40 | — |
DeepSeek V4 Flash DeepSeek | $0.22 | $0.0070 | $0.66 | — |
DeepSeek V4 Pro DeepSeek | $1.32 | $0.04 | $3.96 | — |
Qwen 3.8 Max Qwen | $2 | $0.25 | $6 | — |
Qwen 3.7 Plus Qwen | $0.40 | $0.08 | $1.60 | — |
NVIDIA Nemotron 3.5 Lightning 30B Nemotron | $0.05 | $0.01 | $0.20 | — |
Kimi K3 Kimi | $3 | $0.30 | $15 | — |
MiniMax M3 MiniMax | $0.30 | $0.06 | $1.20 | — |
Kimi K2.7 Code Kimi | $0.95 | $0.19 | $4 | — |
Kimi K2.6 Kimi | $0.95 | $0.16 | $4 | — |
Fireworks AI totals · 1M in + 1M out
NVIDIA Nemotron 3.5 Lightning 30B
$0.2500
GLM 5.3 Flash
$0.6500
OpenAI GPT OSS 120B
$0.7500
DeepSeek V4 Flash
$0.8800
MiniMax M3
$1.50
Qwen 3.7 Plus
$2.00
Kimi K2.7 Code
$4.95
DeepSeek V4 Pro
$5.28
GLM 5.3
$5.80
GLM 5.2
$5.80
Qwen 3.8 Max
$8.00
Kimi K3
$18.00
Cost of 1M input + 1M output tokens. Bar length is square-root scaled so cheap models stay visible next to premium ones.