Kimi K3 pricing and routing profile
Kimi K3 from Moonshot AI lists at $3.00 per million input tokens and $15.00 per million output tokens, with a 1,048,576-token context window. Prices from OpenRouter's catalogue on 2026-09-28.
Tool callingStructured outputReasoning controlLog probabilities
Cost per 1,000 requests
| Request shape | List cost |
|---|---|
| Classification or routing (600 tokens in, 10 out) | $1.95 |
| Chat reply (2,000 in, 400 out) | $12.00 |
| Retrieval-augmented answer (8,000 in, 500 out) | $31.50 |
Before prompt caching. Try your own shape in the cost calculator.
Can Gimbix route to it first?
Yes. 9 of 19 tracked hosts return token log probabilities for Kimi K3, which is the confidence signal Gimbix uses to let a cheaper model answer first.
Who serves Kimi K3
| Host | Input / output per 1M | Log probabilities | Uptime, last 30 min |
|---|---|---|---|
| InferenceNet | $1.00 / $9.00 | no | 99.9% |
| Wafer | $1.00 / $9.00 | no | 100.0% |
| Sail Research | $1.03 / $9.04 | yes | 99.9% |
| Morph | $1.50 / $10.50 | yes | 100.0% |
| Makora | $2.04 / $12.75 | yes | 100.0% |
| Relace | $2.10 / $10.50 | no | 99.8% |
| Phala | $2.55 / $12.75 | yes | 100.0% |
| DigitalOcean | $2.55 / $12.95 | yes | — |
| DeepInfra | $2.85 / $14.25 | no | 99.8% |
| Chutes | $3.00 / $15.00 | no | — |
| Parasail | $3.00 / $15.00 | yes | 98.9% |
| Modal | $3.00 / $15.00 | no | 99.9% |
| Together | $3.00 / $15.00 | no | 99.6% |
| Fireworks | $3.00 / $15.00 | yes | 99.5% |
| BaseTen | $3.00 / $15.00 | no | 99.2% |
| Moonshot AI | $3.00 / $15.00 | no | 99.9% |
| Fireworks | $3.30 / $16.50 | yes | 98.8% |
| Alibaba | $3.45 / $17.25 | yes | 94.0% |
| Fireworks | $4.50 / $22.50 | yes | 100.0% |
A host that lists log probabilities may still omit them; Gimbix checks each one before relying on it.
Cheaper models that can answer first
Models priced below Kimi K3 that expose a confidence signal. Whether they answer your traffic as well is what a Gimbix pilot measures.
| Model | Input / output per 1M | Cheaper on a chat reply |
|---|---|---|
| GLM 5.3 Prime | $2.80 / $8.80 | 24% |
| GPT Audio | $2.50 / $10.00 | 25% |
| GPT-4o (2024-11-20) | $2.50 / $10.00 | 25% |
| GPT-4o (2024-08-06) | $2.50 / $10.00 | 25% |
| GPT-4o | $2.50 / $10.00 | 25% |
| Qwen3.8 Max (0902) | $2.00 / $6.00 | 47% |
Compare Kimi K3
- Kimi K3 vs Claude Opus 5.5
- Kimi K3 vs Claude Sonnet 5
- Kimi K3 vs Claude Fable 5.1
- Kimi K3 vs Claude Opus 5
- Kimi K3 vs GPT-6 Sol
- Kimi K3 vs GPT-6 Astra
- Kimi K3 vs GPT-5.6 Sol
- Kimi K3 vs GPT-5.6 Terra
- Kimi K3 vs GPT-5.5
- Kimi K3 vs GPT-5.4
- Kimi K3 vs Gemini 3.1 Pro Preview
- Kimi K3 vs Gemini 3.8 Flash
- Kimi K3 vs Grok 4.7
- Kimi K3 vs Qwen3.8 Max (0902)
- Kimi K3 vs GLM 5.3
- Kimi K3 vs DeepSeek V4 Pro 0423
- Kimi K3 vs Mistral Large 3 2512
- Kimi K3 vs Nova Premier 1.0
- Kimi K3 vs Command A
- Kimi K3 vs GPT-6 Luna
- Kimi K3 vs GPT-5.6 Luna
- Kimi K3 vs GPT-5.4 Mini
- Kimi K3 vs GPT-5.4 Nano
- Kimi K3 vs GPT-4.1 Mini
- Kimi K3 vs gpt-oss-120b
- Kimi K3 vs gpt-oss-20b
- Kimi K3 vs Claude Haiku 4.5
- Kimi K3 vs Gemini 3.5 Flash Lite
- Kimi K3 vs Gemma 4 31B
- Kimi K3 vs DeepSeek V4.1 Flash
- Kimi K3 vs DeepSeek V4 Flash 0731
- Kimi K3 vs GLM 5.3 Flash
- Kimi K3 vs Hy4 preview
- Kimi K3 vs MiMo-V2.6-Flash
- Kimi K3 vs Nemotron 3.5 Lightning
- Kimi K3 vs Nemotron 3 Ultra
- Kimi K3 vs Qwen3.8 Flash
- Kimi K3 vs Qwen3.7 Flash
- Kimi K3 vs Qwen3.8 27B
- Kimi K3 vs MiniMax M3
- Kimi K3 vs Kimi K2.6
- Kimi K3 vs Mistral Small 4
- Kimi K3 vs Ministral 3 8B 2512
- Kimi K3 vs Llama 4 Maverick
- Kimi K3 vs Llama 3.3 70B Instruct
- Kimi K3 vs Nova Lite 1.0
- Kimi K3 vs Nova Micro 1.0
- Kimi K3 vs Command R7B (12-2024)
- Kimi K3 vs Phi 4
Questions
How much does Kimi K3 cost?
Kimi K3 lists at $3.00 per million input tokens and $15.00 per million output tokens on OpenRouter as of 2026-09-28. A typical chat reply of 2,000 input and 400 output tokens costs about $12.00 per 1,000 requests.
What is the context window of Kimi K3?
1,048,576 tokens.
Can a cheaper model replace Kimi K3?
Sometimes. Gimbix proves it on your own traffic before routing: the cheaper model must match Kimi K3 on one half of your requests and make at most three answers worse on the other.