Kimi K3 pricing & benchmarks
Released Jul 16, 2026 · served by 6 providers · ranked #4 by intelligence in our index
What Kimi K3 is
Kimi K3 is a 2.8T parameter open-weight multimodal reasoning model from Moonshot AI. It is suited for complex coding, knowledge work, and long-horizon agentic workflows, and is particularly strong at...
- Model ID:
moonshotai/kimi-k3 - Modality: text+image->text
- Tool calling: supported · Structured output: supported
- Weights:
moonshotai/Kimi-K3on Hugging Face
Providers and prices for Kimi K3
The same model costs different amounts depending on who serves it. Prices are per 1M tokens, cheapest first. Uptime and throughput are OpenRouter's rolling measurements, not vendor claims.
| Provider | Input | Output | Context | Quant | Uptime 24h | Throughput |
|---|---|---|---|---|---|---|
| BaseTencheapest | $3.00 | $15.00 | 1.0M | fp8 | 97.7% | — |
| DigitalOcean | $3.00 | $15.00 | 1.0M | — | 91.3% | — |
| Nebius | $3.00 | $15.00 | 8K | fp4 | 72.0% | — |
| Together | $3.00 | $15.00 | 1M | — | 89.7% | — |
| Moonshot AI | $3.00 | $15.00 | 1.0M | mxfp4 | 100.0% | — |
| Fireworks | $4.50 | $22.50 | 1.0M | — | 97.6% | — |
Kimi K3 benchmark results
Design Arena head-to-head results, as reported through the OpenRouter model API.
| Category | Arena | Rank | Elo | Win rate |
|---|---|---|---|---|
| 3d | models | #1 | 1455 | 69.3% |
| codecategories | models | #1 | 1420 | 67.3% |
| python-pptxslides | agents | #1 | 1381 | 71% |
| uicomponent | models | #1 | 1414 | 68% |
| webapps | agents | #1 | 1325 | 62.4% |
| website | models | #1 | 1377 | 63.8% |
| dataviz | models | #2 | 1383 | 66.8% |
| godotgamedev | agents | #11 | 1199 | 48.5% |
Cheaper models in the same class
Models scoring within 4 points of Kimi K3 on the Intelligence Index, but with a lower output price.
Compare Kimi K3 head-to-head
FAQ
How much does the Kimi K3 API cost?
$3.00 per 1M input tokens and $15.00 per 1M output tokens. A typical workload of 1M input + 300K output tokens costs about $7.50.
Is Kimi K3 free?
No. It is a paid model starting at $3.00 per 1M input tokens, though some providers offer trial credits.
Which provider is cheapest for Kimi K3?
BaseTen at $3.00 input / $15.00 output per 1M tokens (fp8 quantization).
Can I self-host Kimi K3?
Yes — weights are published as moonshotai/Kimi-K3 on Hugging Face, so you can run it on your own hardware.