MoonshotAI: Kimi K3
by Moonshot AIMoonshotAI: Kimi K3 is a large language model from Moonshot AI. It costs $0.239 per million input tokens and $10.80 per million output tokens. Its context window is 1.0M tokens.
- Input / 1M tokens
- $0.239
- Output / 1M tokens
- $10.80
- Cached input / 1M
- $0.210
- Context window
- 1.0M
On repeated prefixes
tokens
Who serves it cheapest
21 hosts serve Kimi K3. Same weights, same API - the price difference is pure margin and routing.
| Provider | Input / 1M | Output / 1M | Context | Throughput | Uptime 24h |
|---|---|---|---|---|---|
| RelaceCheapestfp4 | $0.182 | $9.34 | 1.0M | - | 99.6% |
| InferenceNetfp4 | $0.239 | $10.80 | 1.0M | - | 100.0% |
| Sail Researchfp4 | $0.230 | $11.00 | 1.0M | - | 100.0% |
| Wafermxfp4 | $0.300 | $10.80 | 1.0M | - | 100.0% |
| Decartmxfp4 | $1.65 | $8.25 | 1.0M | - | 99.6% |
| Makorafp4 | $0.500 | $13.00 | 1.0M | - | 99.9% |
| Chutesmxfp4 | $1.85 | $9.50 | 1.0M | - | 96.4% |
| Fireworks | $2.10 | $10.50 | 1.0M | - | 97.7% |
| AkashMLfp4 | $1.20 | $14.00 | 1.0M | - | 97.2% |
| Morphfp8 | $1.28 | $14.63 | 1.0M | - | 99.9% |
| Phala | $2.55 | $12.75 | 1.0M | - | 99.8% |
| DigitalOcean | $2.55 | $12.95 | 1.0M | - | 100.0% |
| DeepInframxfp4 | $2.85 | $14.25 | 1.0M | - | 100.0% |
| ParetoAlpha | $3.00 | $15.00 | 1.0M | - | 100.0% |
| Amazon Bedrock | $3.00 | $15.00 | 1.0M | - | 97.9% |
| Parasailfp4 | $3.00 | $15.00 | 1.0M | - | 99.6% |
| Modalmxfp4 | $3.00 | $15.00 | 1.0M | - | 100.0% |
| Together | $3.00 | $15.00 | 1.0M | - | 99.7% |
| BaseTenfp8 | $3.00 | $15.00 | 1.0M | - | 100.0% |
| Moonshot AI | $3.00 | $15.00 | 1.0M | - | 100.0% |
| Alibaba | $3.45 | $17.25 | 1.0M | - | 96.3% |
The spread between Relace and Alibaba is 2.8× for identical weights. Quantization and context limits differ, so check both columns before switching.
Benchmarks
Independent scores published alongside the catalogue.
- Coding index
- 76.2
- Agentic index
- 50.0
About Kimi K3
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...
Specifications
| Model ID | moonshotai/kimi-k3 |
|---|---|
| Provider | Moonshot AI |
| Context window | 1.0M tokens |
| Max output | 944K tokens |
| Input modalities | text, image, video |
| Output modalities | text |
| Knowledge cutoff | - |
| Open weights | Yes - moonshotai/Kimi-K3 |
| Released | July 16, 2026 |
Cheaper alternatives
Models that cost less than Kimi K3 while keeping at least half its context window and every input modality it supports.
Frequently asked
How much does MoonshotAI: Kimi K3 cost?
$0.239 per million input tokens and $10.80 per million output tokens. Cached input reads cost $0.210 per million tokens.
What is the context window of MoonshotAI: Kimi K3?
1.0M tokens, with up to 944K tokens of output per request.
Which provider serves MoonshotAI: Kimi K3 cheapest?
Relace at $0.182 per million input tokens - 2.8× cheaper than Alibaba, the most expensive of the 21 hosts serving it.
Confirm against the source: Moonshot AI official pricing.