MoonshotAI: Kimi K3
6 providers live- 6 live providers
- 100% uptime
- 0.83 s first token
- 78 tok/s output
- 10% off
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 navigating large repositories, using tools, debugging, and iterating against images, logs, tests, and runtime feedback. Its architecture uses KDA and Attention Residuals for computational efficiency.
- Context
- 1M
- tokens
- Max output
- 943.7K
- tokens
- Input
$2.85$2.565- per million tokens, 10% off
- Output
$14.25$12.825- per million tokens, 10% off
Providers
Ranked by price and speed. Requests go to the best healthy provider, then fail over down the list.Benchmarks| Who answers now | Input per 1M | First token |
|---|---|---|
 llm.buzz10% offpicked | 4.29 s | |
 DeepInfra | $2.85 | 0.83 s |
| Moonshot AI | $3.00 | 1.44 s |
 Prime Intellect | $3.45 | 0.83 s |
 Novita | $3.00 | 2.98 s |
 Alibaba | $3.45 | 1.44 s |
llm.buzz answers now: lowest price.
First token is the 24-hour median. If the pick fails, the next one answers at its own price.
| # | Provider | Input | Output | Cache read | Latency | Throughput | Uptime (3d) |
|---|---|---|---|---|---|---|---|
| 1 | Data policy not declared | 4.29 s | 54 tok/s | 98.6% | |||
| 2 | Data policy not declared | $2.85 | $14.25 | $0.285 | 0.83 s | 78 tok/s | 100% |
| 3 | Data policy not declared | $3.00 | $15.00 | $0.30 | 1.44 s | 65 tok/s | 100% |
| 4 | Data policy not declared | $3.45 | $17.25 | n/a | 0.83 s | 65 tok/s | 98.6% |
| 5 | Data policy not declared | $3.00 | $15.00 | $0.30 | 2.98 s | 54 tok/s | 100% |
| 6 | Data policy not declared | $3.45 | $17.25 | $0.345 | 1.44 s | 54 tok/s | 100% |
Price history
No price changes since Oct 20, 2026. The history shows prices before discounts.
- DeepInfra$2.85 in, $14.25 out
- Alibaba$3.45 in, $17.25 out
- Prime Intellect$3.45 in, $17.25 out
- Novita$3.00 in, $15.00 out
- Moonshot AI$3.00 in, $15.00 out
Performance
Last 24 hours, from llm.buzz's test requestsThroughput (tok/s)
- llm.buzz54 tok/s
- DeepInfra78 tok/s
- Moonshot AI65 tok/s
- Prime Intellect65 tok/s
- Novita54 tok/s
- Alibaba54 tok/s
Data table
| Provider | From (UTC) | p50 | p90 | p99 |
|---|---|---|---|---|
| llm.buzz | Oct 26, 07:00 | 54 tok/s | 54 tok/s | 54 tok/s |
| llm.buzz | Oct 26, 12:00 | 65 tok/s | 65 tok/s | 65 tok/s |
| llm.buzz | Oct 26, 16:00 | 37 tok/s | 37 tok/s | 37 tok/s |
| llm.buzz | Oct 26, 21:00 | 54 tok/s | 54 tok/s | 54 tok/s |
| llm.buzz | Oct 27, 01:00 | 78 tok/s | 78 tok/s | 78 tok/s |
| llm.buzz | Last 24 hours | 54 tok/s | 37 tok/s | 37 tok/s |
| DeepInfra | Oct 26, 06:00 | 54 tok/s | 54 tok/s | 54 tok/s |
| DeepInfra | Oct 26, 10:00 | 93 tok/s | 93 tok/s | 93 tok/s |
| DeepInfra | Oct 26, 14:00 | 78 tok/s | 78 tok/s | 78 tok/s |
| DeepInfra | Oct 26, 19:00 | 78 tok/s | 78 tok/s | 78 tok/s |
| DeepInfra | Oct 27, 00:00 | 26 tok/s | 26 tok/s | 26 tok/s |
| DeepInfra | Last 24 hours | 78 tok/s | 26 tok/s | 26 tok/s |
| Moonshot AI | Oct 26, 04:00 | 54 tok/s | 54 tok/s | 54 tok/s |
| Moonshot AI | Oct 26, 09:00 | 54 tok/s | 54 tok/s | 54 tok/s |
| Moonshot AI | Oct 26, 14:00 | 65 tok/s | 65 tok/s | 65 tok/s |
| Moonshot AI | Oct 26, 19:00 | 65 tok/s | 65 tok/s | 65 tok/s |
| Moonshot AI | Oct 27, 00:00 | 93 tok/s | 93 tok/s | 93 tok/s |
| Moonshot AI | Last 24 hours | 65 tok/s | 54 tok/s | 54 tok/s |
| Prime Intellect | Oct 26, 06:00 | 54 tok/s | 54 tok/s | 54 tok/s |
| Prime Intellect | Oct 26, 10:00 | 78 tok/s | 78 tok/s | 78 tok/s |
| Prime Intellect | Oct 26, 15:00 | 54 tok/s | 54 tok/s | 54 tok/s |
| Prime Intellect | Oct 26, 19:00 | 65 tok/s | 65 tok/s | 65 tok/s |
| Prime Intellect | Oct 27, 00:00 | 78 tok/s | 78 tok/s | 78 tok/s |
| Prime Intellect | Last 24 hours | 65 tok/s | 54 tok/s | 54 tok/s |
| Novita | Oct 26, 04:00 | 65 tok/s | 65 tok/s | 65 tok/s |
| Novita | Oct 26, 09:00 | 54 tok/s | 54 tok/s | 54 tok/s |
| Novita | Oct 26, 13:00 | 45 tok/s | 45 tok/s | 45 tok/s |
| Novita | Oct 26, 18:00 | 65 tok/s | 65 tok/s | 65 tok/s |
| Novita | Oct 26, 22:00 | 54 tok/s | 54 tok/s | 54 tok/s |
| Novita | Oct 27, 02:00 | 54 tok/s | 54 tok/s | 54 tok/s |
| Novita | Last 24 hours | 54 tok/s | 45 tok/s | 45 tok/s |
| Alibaba | Oct 26, 05:00 | 54 tok/s | 54 tok/s | 54 tok/s |
| Alibaba | Oct 26, 10:00 | 45 tok/s | 45 tok/s | 45 tok/s |
| Alibaba | Oct 26, 15:00 | 45 tok/s | 45 tok/s | 45 tok/s |
| Alibaba | Oct 26, 19:00 | 54 tok/s | 54 tok/s | 54 tok/s |
| Alibaba | Oct 26, 23:00 | 54 tok/s | 54 tok/s | 54 tok/s |
| Alibaba | Last 24 hours | 54 tok/s | 45 tok/s | 45 tok/s |
Latency (time to first token)
- llm.buzz4.29 s
- DeepInfra0.83 s
- Moonshot AI1.44 s
- Prime Intellect0.83 s
- Novita2.98 s
- Alibaba1.44 s
Data table
| Provider | From (UTC) | p50 | p90 | p99 |
|---|---|---|---|---|
| llm.buzz | Oct 26, 07:00 | 2.07 s | 2.07 s | 2.07 s |
| llm.buzz | Oct 26, 12:00 | 38.2 s | 38.2 s | 38.2 s |
| llm.buzz | Oct 26, 16:00 | 55.0 s | 55.0 s | 55.0 s |
| llm.buzz | Oct 26, 21:00 | 2.98 s | 2.98 s | 2.98 s |
| llm.buzz | Oct 27, 01:00 | 4.29 s | 4.29 s | 4.29 s |
| llm.buzz | Last 24 hours | 4.29 s | 55.0 s | 55.0 s |
| DeepInfra | Oct 26, 06:00 | 0.83 s | 0.83 s | 0.83 s |
| DeepInfra | Oct 26, 10:00 | 0.69 s | 0.69 s | 0.69 s |
| DeepInfra | Oct 26, 14:00 | 0.83 s | 0.83 s | 0.83 s |
| DeepInfra | Oct 26, 19:00 | 0.83 s | 0.83 s | 0.83 s |
| DeepInfra | Oct 27, 00:00 | 0.48 s | 0.48 s | 0.48 s |
| DeepInfra | Last 24 hours | 0.83 s | 0.83 s | 0.83 s |
| Moonshot AI | Oct 26, 04:00 | 1.20 s | 1.20 s | 1.20 s |
| Moonshot AI | Oct 26, 09:00 | 2.07 s | 2.07 s | 2.07 s |
| Moonshot AI | Oct 26, 14:00 | 2.48 s | 2.48 s | 2.48 s |
| Moonshot AI | Oct 26, 19:00 | 1.44 s | 1.44 s | 1.44 s |
| Moonshot AI | Oct 27, 00:00 | 1.44 s | 1.44 s | 1.44 s |
| Moonshot AI | Last 24 hours | 1.44 s | 2.48 s | 2.48 s |
| Prime Intellect | Oct 26, 06:00 | 0.83 s | 0.83 s | 0.83 s |
| Prime Intellect | Oct 26, 10:00 | 0.83 s | 0.83 s | 0.83 s |
| Prime Intellect | Oct 26, 15:00 | 1.00 s | 1.00 s | 1.00 s |
| Prime Intellect | Oct 26, 19:00 | 1.72 s | 1.72 s | 1.72 s |
| Prime Intellect | Oct 27, 00:00 | 0.83 s | 0.83 s | 0.83 s |
| Prime Intellect | Last 24 hours | 0.83 s | 1.72 s | 1.72 s |
| Novita | Oct 26, 04:00 | 10.7 s | 10.7 s | 10.7 s |
| Novita | Oct 26, 09:00 | 2.48 s | 2.48 s | 2.48 s |
| Novita | Oct 26, 13:00 | 6.17 s | 6.17 s | 6.17 s |
| Novita | Oct 26, 18:00 | 10.7 s | 10.7 s | 10.7 s |
| Novita | Oct 26, 22:00 | 2.98 s | 2.98 s | 2.98 s |
| Novita | Oct 27, 02:00 | 2.98 s | 2.98 s | 2.98 s |
| Novita | Last 24 hours | 2.98 s | 10.7 s | 10.7 s |
| Alibaba | Oct 26, 05:00 | 1.44 s | 1.44 s | 1.44 s |
| Alibaba | Oct 26, 10:00 | 1.00 s | 1.00 s | 1.00 s |
| Alibaba | Oct 26, 15:00 | 1.20 s | 1.20 s | 1.20 s |
| Alibaba | Oct 26, 19:00 | 2.07 s | 2.07 s | 2.07 s |
| Alibaba | Oct 26, 23:00 | 1.72 s | 1.72 s | 1.72 s |
| Alibaba | Last 24 hours | 1.44 s | 2.07 s | 2.07 s |
Uptime
Last 72 hours, by hour- Up: 95% or more served
- Degraded: 85% to 95%
- Down: below 85%
- No data
- From llm.buzz's test requests
Questions
How much does MoonshotAI: Kimi K3 cost?
MoonshotAI: Kimi K3 costs $2.565 per million input tokens and $12.825 per million output tokens at llm.buzz, the provider requests go to first, with cache reads at $0.2565 per million tokens. These prices are 10% off until 2026-11-06 00:00 UTC. Each provider sets its own price, and you pay the price of the provider that answered.
Do providers train on my prompts?
Each provider declares it, and the table above shows every declaration. You can limit a request to providers that keep nothing, or to providers in the EU.
Can I serve MoonshotAI: Kimi K3 myself?
Yes. Apply as a provider, add your endpoint and list this model at your price. It goes live once the automated check passes.
Similar models
Call it
import os
from openai import OpenAI
client = OpenAI(
base_url="https://api.llm.buzz/api/v1",
api_key=os.environ["LLM_BUZZ_KEY"],
)
reply = client.chat.completions.create(
model="moonshotai/kimi-k3",
messages=[{"role": "user", "content": "Hello"}],
)import OpenAI from "openai";
const client = new OpenAI({
baseURL: "https://api.llm.buzz/api/v1",
apiKey: process.env.LLM_BUZZ_KEY,
});
const reply = await client.chat.completions.create({
model: "moonshotai/kimi-k3",
messages: [{ role: "user", content: "Hello" }],
});curl https://api.llm.buzz/api/v1/chat/completions \
-H "Authorization: Bearer $LLM_BUZZ_KEY" \
-H "Content-Type: application/json" \
-d '{"model": "moonshotai/kimi-k3",
"messages": [{"role": "user", "content": "Hello"}]}'Endpoints
POST /api/v1/chat/completionsPOST /api/v1/responsesPOST /api/v1/messagesAuthentication
Send your llm.buzz key as a Bearer token: Authorization: Bearer $LLM_BUZZ_KEY. Create keys in the console. The same key calls every model; each request pays the price of the provider that answered it.
More in Authentication and the API reference.
README
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 navigating large repositories, using tools, debugging, and iterating against images, logs, tests, and runtime feedback. Its architecture uses KDA and Attention Residuals for computational efficiency.
- Model ID
- moonshotai/kimi-k3
- Takes
- Text, image
- Answers
- Text
- Context
- 1M tokens
- Max output
- 943.7K tokens
- Capabilities
- Tools, Caching, Structured outputs
- Parameters
- max_tokens, response_format, structured_outputs, temperature, tool_choice, tools
- Weights
- Open weights
- On llm.buzz since
- Oct 20, 2026