Kimi K3
- Input / output
- $0.276 / $9.60
- Providers
- 20
- Spread
- 2.3×
- 7 days
- 0%
- 30 days
- +93.1%
- Context
- 1M
Providers
| Provider | Input | Output | Blended | vs ref | Context | Uptime 1d | Price since |
|---|---|---|---|---|---|---|---|
| mxfp4 | $0.276 | $9.60 | $2.607 | −32% | |||
| fp4 | $0.181 | $11.01 | $2.89 | −25% | |||
| mxfp4 | $1.85 | $9.50 | $3.76 | −2.2% | |||
| fp4 | $0.245 | $15 | $3.93 | +2.2% | |||
| fp4 | $0.245 | $15 | $3.93 | +2.2% | |||
| $0.45 | $15 | $4.09 | +6.2% | ||||
Waferfast mxfp4 | $0.19 | $16 | $4.14 | +7.6% | |||
| mxfp4 | $0.192 | $16 | $4.144 | +7.7% | |||
| fp4 | $0.345 | $16.58 | $4.40 | +14% | |||
InferenceNetfast fp4 | $0.45 | $18 | $4.84 | +26% | |||
InferenceNetus/fast fp4 | $0.45 | $18 | $4.84 | +26% | |||
| fp4 | $1.90 | $14 | $4.925 | +28% | |||
| $2.55 | $12.95 | $5.15 | +34% | ||||
| fp4 | $1.00 | $18 | $5.25 | +36% | |||
| mxfp4 | $2.85 | $14.25 | $5.70 | +48% | |||
| $3.00 | $15 | $6.00 | +56% | ||||
Amazon Bedrockus-east-2 | $3.00 | $15 | $6.00 | +56% | |||
| fp8 | $3.00 | $15 | $6.00 | +56% | |||
| $3.00 | $15 | $6.00 | +56% | ||||
| mxfp4 | $3.00 | $15 | $6.00 | +56% | |||
| $3.00 | $15 | $6.00 | +56% | ||||
| fp4 | $3.00 | $15 | $6.00 | +56% | |||
| fp4 | $3.00 | $15 | $6.00 | +56% | |||
| $3.00 | $15 | $6.00 | +56% | ||||
| $3.00 | $15 | $6.00 | +56% | ||||
Amazon Bedrockus-east-1 | $3.30 | $16.50 | $6.60 | +72% | |||
Fireworksfast | $4.50 | $22.50 | $9.00 | +134% | |||
| $4.50 | $22.50 | $9.00 | +134% | ||||
Parasailfast fp4 | $4.50 | $22.50 | $9.00 | +134% |
- WaferCheapestus$2.607$0.276 in · $9.60 out−32% vs ref1M ctx100% uptimemxfp4
- $2.89$0.181 in · $11.01 out−25% vs ref1M ctx99.9% uptimefp4
- $3.76$1.85 in · $9.50 out−2.2% vs ref1M ctx95.8% uptimemxfp4
- $3.93$0.245 in · $15 out+2.2% vs ref1M ctx100% uptimefp4
- $3.93$0.245 in · $15 out+2.2% vs ref1M ctx100% uptimefp4
- $4.09$0.45 in · $15 out+6.2% vs ref1M ctx99.2% uptime
- fast$4.14$0.19 in · $16 out+7.6% vs ref1M ctx100% uptimemxfp4
- $4.144$0.192 in · $16 out+7.7% vs ref1M ctx100% uptimemxfp4
- $4.40$0.345 in · $16.58 out+14% vs ref1M ctx100% uptimefp4
- fast$4.84$0.45 in · $18 out+26% vs ref1M ctx100% uptimefp4
- us/fast$4.84$0.45 in · $18 out+26% vs ref1M ctx100% uptimefp4
- $4.925$1.90 in · $14 out+28% vs ref1M ctx99.1% uptimefp4
- $5.15$2.55 in · $12.95 out+34% vs ref1M ctx100% uptime
- $5.25$1.00 in · $18 out+36% vs ref1M ctx99.9% uptimefp4
- $5.70$2.85 in · $14.25 out+48% vs ref1M ctx99.8% uptimemxfp4
- $6.00$3.00 in · $15 out+56% vs ref1M ctx100% uptime
- us-east-2$6.00$3.00 in · $15 out+56% vs ref1M ctx100% uptime
- $6.00$3.00 in · $15 out+56% vs ref1M ctx99.7% uptimefp8
- $6.00$3.00 in · $15 out+56% vs ref1M ctx99.7% uptime
- $6.00$3.00 in · $15 out+56% vs ref1M ctx100% uptimemxfp4
- $6.00$3.00 in · $15 out+56% vs ref1M ctx99.4% uptime
- $6.00$3.00 in · $15 out+56% vs ref1M ctx100% uptimefp4
- $6.00$3.00 in · $15 out+56% vs ref1M ctx100% uptimefp4
- $6.00$3.00 in · $15 out+56% vs ref1M ctx99.6% uptime
- $6.00$3.00 in · $15 out+56% vs ref1M ctx98.4% uptime
- us-east-1$6.60$3.30 in · $16.50 out+72% vs ref1M ctx100% uptime
- fast$9.00$4.50 in · $22.50 out+134% vs ref1M ctx98.7% uptime
- $9.00$4.50 in · $22.50 out+134% vs ref1M ctx97.2% uptime
- fast$9.00$4.50 in · $22.50 out+134% vs ref1M ctx100% uptimefp4
Blended = (3 × input + output) / 4. Ref $3.85 = median of cheapest 20%.
Price history
USD per 1M tokens · each step is a price change
Prices in USD / 1M tokens.
Other pricing
| Offered by | Cache read | Cache write |
|---|---|---|
| Wafer (us) | $0.13 | — |
| Relace | $0.18 | — |
| Chutes | $0.185 | — |
| InferenceNet, InferenceNet (us) | $0.214 | — |
| KT, Alibaba +8 | $0.30 | — |
| Wafer (fast) | $0.152 | — |
| Wafer | $0.182 | — |
| Sail Research | $0.173 | — |
| InferenceNet (fast), InferenceNet (us/fast) | $0.33 | — |
| Io Net | $0.19 | — |
| DigitalOcean | $0.255 | — |
| Makora | $0.50 | — |
| DeepInfra | $0.285 | — |
| Amazon Bedrock (us-east-2) | $0.30 | $3.75 |
| Amazon Bedrock (us-east-1) | $0.33 | $4.125 |
| Fireworks (fast), Fireworks (us) +1 | $0.45 | — |
Changes
- PriceInput price,$0.153→$0.19+24.2%
- PriceInput price,$0.182→$0.181−1.0%
- PriceCache read price,$0.204→$0.214+4.9%
- PriceInput price,$0.234→$0.245+4.7%
- PriceCache read price,$0.204→$0.214+4.9%
- PriceInput price,$0.234→$0.245+4.7%
- PriceInput price,$1.20→$1.00−16.7%
- ListedNow served,Listed
- PriceInput price,$1.00→$1.20+20.0%
- ListedNow served,Listed
- PriceCache read price,$0.315→$0.33+4.8%
- PriceCache read price,$0.315→$0.33+4.8%
- PriceCache read price,$0.21→$0.19−9.5%
- PriceInput price,$2.10→$1.90−9.5%
- DelistedNo longer served,Delisted
About
The cheapest provider for Kimi K3 is Wafer at $2.607 per 1M tokens (blended), charging $0.276 per 1M input and $9.60 per 1M output tokens. 20 providers serve it; the most expensive charges 2.3× the typical price. Its best price rose 93.1% over the last 30 days. Context window: 1M tokens. Prices tracked since Oct 3, 2026.
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...
- Released
- Jul 16, 2026
- First tracked
- Oct 3, 2026
- Canonical slug
- moonshotai/kimi-k3-20260715
- Weights
- moonshotai/Kimi-K3