Moonshot AI · open weights
Kimi K3 API pricing by provider
Kimi K3 is available from 21 providers. The cheapest is Relace at $0.17 input / $7.697 output per 1M tokens (FP4). Moonshot AI’s own API charges $3.00 / $15.00, 66% more than the cheapest host. The priciest host, Fireworks, charges 4.4x the cheapest.
Prices from the OpenRouter API, retrieved October 26, 2026. Weights: moonshotai/Kimi-K3.
| Provider | Precision | Input / 1M | Output / 1M | Document Q&A / month | Context | Max output |
|---|---|---|---|---|---|---|
| Relace | fp4 | $0.17 | $7.697 | $298.92 | 1.05M | 944K |
| Sail Research | fp4 | $1.20 | $5.50 | $645.00 | 1.05M | 944K |
| InferenceNet | fp4 | $0.239 | $10.80 | $419.60 | 1.05M | 944K |
| Wafer | mxfp4 | $0.36 | $11.90 | $501.00 | 1.05M | 944K |
| Decart | mxfp4 | $1.65 | $8.25 | $907.50 | 1.05M | 944K |
| InferenceNet | fp4 | $0.50 | $11.88 | $556.40 | 1.05M | 944K |
| Chutes | mxfp4 | $1.85 | $9.50 | $1,025 | 1.05M | 66K |
| Fireworks | — | $2.10 | $10.50 | $1,155 | 1.05M | 944K |
| Makora | fp4 | $0.65 | $15.00 | $710.00 | 1.05M | 944K |
| AkashML | fp4 | $1.20 | $14.00 | $900.00 | 1.05M | 944K |
| Phala | — | $2.55 | $12.75 | $1,403 | 1.05M | 944K |
| Wafer | mxfp4 | $0.179 | $20.00 | $671.60 | 1.05M | 944K |
| DigitalOcean | — | $2.55 | $12.95 | $1,409 | 1.05M | 944K |
| InferenceNet | fp4 | $2.24 | $14.52 | $1,332 | 1.05M | 944K |
| DeepInfra | mxfp4 | $2.85 | $14.25 | $1,568 | 1.05M | 944K |
| ParetoAlpha | — | $3.00 | $15.00 | $1,650 | 1.05M | 944K |
| Amazon Bedrock | — | $3.00 | $15.00 | $1,650 | 1.05M | 128K |
| Parasail | fp4 | $3.00 | $15.00 | $1,650 | 1.05M | 944K |
| Modal | mxfp4 | $3.00 | $15.00 | $1,650 | 1.05M | 944K |
| Together | — | $3.00 | $15.00 | $1,650 | 1.05M | 944K |
| BaseTen | fp8 | $3.00 | $15.00 | $1,650 | 1.05M | 262K |
| Moonshot AIcreator | mxfp4 | $3.00 | $15.00 | $1,650 | 1.05M | 944K |
| Alibaba | — | $3.45 | $17.25 | $1,897 | 1.05M | 131K |
| Morph | fp8 | $1.384 | $29.79 | $1,447 | 1.05M | 944K |
| Fireworks | — | $4.50 | $22.50 | $2,475 | 1.05M | 944K |
Sorted by blended price (3 input : 1 output). Document Q&A assumes 50,000 questions a month with 8,000 input and 600 output tokens each. Precision is reported by the provider; “—” means not disclosed.
Frequently asked questions
- What is the cheapest Kimi K3 API provider?
- Relace is the cheapest healthy provider we track at $0.17 per 1M input tokens and $7.697 per 1M output tokens.
- How much does Kimi K3 cost?
- Across 21 providers, input prices range from $0.17 to $4.50 per 1M tokens and output prices from $5.50 to $29.79.
- What is Kimi K3’s context window?
- Kimi K3 supports up to 1.05M tokens, but some providers serve a smaller context window — check the table before choosing.
Other open-weight models
- DeepSeek V4.1 Flashfrom $0.09 / $0.25
- DeepSeek V4 Pro 0813from $0.132 / $0.396
- GLM 5.3from $0.0699 / $1.40
- GLM 5.3 Flashfrom $0.0413 / $0.0758
- Qwen3.8 27Bfrom $0.02 / $0.50
- MiniMax M3from $0.224 / $0.88
- gpt-oss-120bfrom $0.03 / $0.15
- gpt-oss-20bfrom $0.018 / $0.09
- Llama 4 Maverickfrom $0.1875 / $0.6525
- Llama 3.3 70B Instructfrom $0.10 / $0.32
- Gemma 4 31Bfrom $0.10 / $0.34
- Mistral Small 4from $0.15 / $0.60
- Nemotron 3 Superfrom $0.085 / $0.40