Models / Kimi K3 / Calculator

Kimi K3 cost & VRAM calculator

What does Kimi K3 cost to run for your workload, and can you run it on your own hardware? Set your workload below - we compute per-provider API cost live and tell you honestly whether local hardware can run it.

Your workload

Casual Power user Team of 5 Codebase task

Kimi K3 runs an always-on thinking mode. Reasoning (thinking) tokens are billed at the output rate ($15.00/M), so count them here to see the thinking portion of your bill.

API cost for your workload

Provider Rate ($/1M) Monthly cost
Relace $0.23 in · $8.52 out · $0.22 cache $0.01 cheapest
Wafer $0.36 in · $14.90 out · $0.36 cache $0.01 cheapest
Makora $0.65 in · $15.00 out · $0.23 cache $0.02
Morph $1.15 in · $16.09 out · $0.41 cache $0.03
AkashML $1.20 in · $14.00 out · $1.20 cache $0.03
Sail Research $1.20 in · $5.50 out · $0.20 cache $0.03
Decart $1.65 in · $8.25 out · $0.16 cache $0.04
InferenceNet $2.24 in · $14.52 out · $0.35 cache $0.05
Nous Portal may be stale $2.40 in · $12.00 out $0.06
DigitalOcean $2.55 in · $12.95 out · $0.26 cache $0.06
Phala $2.55 in · $12.75 out · $0.26 cache $0.06
DeepInfra $2.85 in · $14.25 out · $0.28 cache $0.07
DigitalOcean may be stale $2.85 in · $14.25 out · $0.28 cache $0.07
BaseTen $3.00 in · $15.00 out · $0.30 cache $0.07
Ollama Cloud may be stale $3.00 in · $15.00 out · $0.30 cache $0.07
Modal $3.00 in · $15.00 out · $0.30 cache $0.07
Moonshot AI $3.00 in · $15.00 out · $0.30 cache $0.07
ParetoAlpha $3.00 in · $15.00 out · $0.30 cache $0.07
Together $3.00 in · $15.00 out · $0.30 cache $0.07
Amazon Bedrock $3.00 in · $15.00 out · $0.30 cache $0.07
Alibaba $3.45 in · $17.25 out · $0.34 cache $0.08
Fireworks $4.50 in · $22.50 out · $0.45 cache $0.11
Parasail $3.00 in · $15.00 out · $0.30 cache $0.07
Chutes avoid $1.85 in · $9.50 out · $0.18 cache $0.04

Monthly cost is an estimate from list prices and your workload - verify against the provider before committing. Cached fraction applies the cache rate to that share of input.

Can you run it locally?

YES - IT RUNS ON HARDWARE YOU CAN BUY

Kimi K3 fits on individual rigs. Here's the hardware that runs it comfortably; the full per-quant VRAM matrix and tok/s estimates live on the model page.

NVIDIA DGX Station 748GB 748GB UD-IQ1_S
199.4t/s
Dual EPYC 9004 + 768GB DDR5-4800 768GB UD-IQ1_S
11.5t/s
UD-IQ1_S
594.0GB weights 610.0GB min 665.0GB rec

Want the own-vs-rent payback (capex vs your monthly API bill)? The budget tool computes that against representative rigs.

Full model card API pricing table Generic token calculator