inference + compute
two ways to buy
Per-token inference with prepaid credits — or dedicated 8-GPU NVIDIA B200 & B300 HGX nodes, prepaid by the term. No subscriptions. No surprises. OpenAI-compatible API for DeepSeek V4.1, MiniMax M3, GLM-5.3, Kimi K3, MiMo-V2.6 and more.
Get Started FreePay per token from your prepaid balance — no minimum spend
Dedicated Deployments
Need guaranteed capacity? Reserve dedicated 8-GPU NVIDIA B200 or B300 HGX nodes with locked rates — 12 months to 3 years, prepaid. See Compute pricing →
All models run on NVIDIA B200 GPUs with speculative decoding for industry-leading speeds. Pay only for the tokens you use with prepaid credits.
Reserved, single-tenant 8-GPU HGX nodes. The whole term is billed once — your rate is locked.
| GPU | 1 Month | 3 Months | 6 Months | 12 Months | 24 Months | 3 YearsBest value |
|---|---|---|---|---|---|---|
|
NVIDIA B200
Blackwell · 180 GB
|
$5.40/hr
Sold out
|
$4.95/hr
Sold out
|
$4.77/hr
Sold out
|
$4.32/hr
$302,746 per node
|
$3.87/hr
$542,419 per node
|
$3.42/hr
$719,021 per node
|
|
NVIDIA B300
Blackwell Ultra · 288 GB
|
$5.40/hr
Sold out
|
$5.25/hr
Sold out
|
$5.15/hr
Sold out
|
$4.95/hr
$346,896 per node
|
$4.59/hr
$643,334 per node
|
$4.23/hr
$889,315 per node
|
All rates are per GPU, per hour, on complete 8-GPU HGX nodes — one node minimum, whole term prepaid upfront (730 hr/GPU/month basis), plus applicable sales tax/VAT where applicable. Example: 8× B300 · 3 years = $889,315 prepaid. Volume pricing for multi-node clusters; custom terms and durations on request. Available in North America (US) only. Prices as of Oct 2026 — reserved rates lock at order.
489 tokens/sec means your AI assistant thinks faster. Cursor autocomplete feels instant, Claude Code edits land quicker, and coding agents iterate in seconds instead of minutes.
Top up your balance and start making API calls immediately
Everything you need to know about Avian pricing
Avian offers 10 models across the DeepSeek, MiniMax, GLM, Kimi and MiMo families: DeepSeek V4.1 Flash, DeepSeek V4 Flash, DeepSeek V4 Pro 0813, MiniMax M3, GLM-5.3, GLM-5.3 Flash, Kimi K3, MiMo-V2.6 Flash, MiMo-V2.6 Pro and MiMo-V2.5 Pro. All models are available to every user — just add credits and start making requests.
Avian uses a simple prepaid credit system. Add credits to your account, then each API request deducts the token cost from your balance. No subscriptions, no monthly fees — you only pay for the tokens you use.
Yes. Reserve dedicated NVIDIA B200 or B300 GPUs — complete 8-GPU HGX nodes, single-tenant, one node minimum — prepaid for a 12-month, 24-month, or 3-year term with your rate locked. See Compute pricing →
Yes. The Avian API follows the OpenAI Chat Completions format. Just change the base URL to https://api.avian.io/v1 in any OpenAI SDK and it works out of the box.
No. Credits never expire. Use them whenever you need — there's no time limit on your prepaid balance.
No. As long as you have credits in your account, you can make as many API requests as you need. Your balance is the only limit.