Requests
10/3011/05 UTC
Requests in the selected window, excluding refund records.
← Back to models
Recorded declaration
View reference source ↗
Get an API key ↗
Kimi-K3
API model IDAutomatic routing
Kimi K3 targets long-horizon coding, knowledge work and deep reasoning. It suits development analysis, synthesis across multiple documents and planning complex workflows.Read moreRead less
Kimi K3 targets long-horizon coding, knowledge work and deep reasoning. It suits development analysis, synthesis across multiple documents and planning complex workflows.
Providers
Different providers serve this model. Compare pricing and performance, select a provider, or use automatic routing.
| Provider | Variant / API model ID | Input / M | Output / M | Cache read / M | Request latency | Output rate | Request success | Connect |
|---|---|---|---|---|---|---|---|---|
Kimi-K3✓ Official routekimi-k3-ad | ¥17 | ¥85 | ¥1.7 | 14.55 s | 7.7 tokens/s | 100.00% | Use route ↗ |
Route details kimi-k3-ad
✓Latency14.55 s
Throughput7.7 tokens/s
Request success100.00%
Input¥17 / 1M tokens
Output¥85 / 1M tokens
Context1,000,000 tokens
Recent request health
100.00%EarlierLatest
Parameters & pricing
Values apply to the currently selected API model ID. kimi-k3-ad
- Context limit
- 1000000 tokens
Declared and manually configured parameters are informational. They do not enable capabilities or change gateway limits; actual requests remain subject to the selected route’s validation and upstream support.
Your applicable prices
| Specification | Sale price | Comparable official price | Official price comparison |
|---|---|---|---|
| Current model | Input CNY ¥17 / 1M tokensOutput CNY ¥85 / 1M tokensCache read CNY ¥1.7 / 1M tokens | Input USD $3.0 / 1M tokensOutput USD $15 / 1M tokensCache read USD $0.30 / 1M tokensCache write USD $3.0 / 1M tokens | Different currencies; discount not calculated |
Discounts require a verified reference and matching billing units and specifications. USD references use the explicitly configured exchange rate; no rate means no conversion.
API integration
- Base URL
- https://api.ai526.com/v1
- Request endpoint
- POST /v1/chat/completions
- model
- kimi-k3-ad
Use your platform API key. A unique model ID fixes this route; a family ID selects the lowest-priced route.
Request parameters · JSON
{
"model": "kimi-k3-ad",
"messages": [
{
"role": "user",
"content": "Write a short description of the future."
}
]
}cURL example
curl https://api.ai526.com/v1/chat/completions \
-H 'Authorization: Bearer YOUR_AI526_API_KEY' \
-H 'Content-Type: application/json' \
-d '{"model":"kimi-k3-ad","messages":[{"role":"user","content":"Write a short description of the future."}]}'
# API URL: use the URL above.
# YOUR_AI526_API_KEY: replace with your own platform key.
# model: use the copied unique ID, or a family ID for the lowest-priced route.
# prompt / messages: replace with your task.Python example
import requests
response = requests.post(
"https://api.ai526.com/v1/chat/completions",
headers={
"Authorization": "Bearer YOUR_AI526_API_KEY",
"Content-Type": "application/json"
},
json={
"model": "kimi-k3-ad",
"messages": [
{
"role": "user",
"content": "Write a short description of the future."
}
]
},
)
response.raise_for_status()
print(response.json())
# API URL: use the URL above.
# YOUR_AI526_API_KEY: replace with your own platform key.
# model: use the copied unique ID, or a family ID for the lowest-priced route.
# prompt / messages: replace with your task.JavaScript example
const response = await fetch("https://api.ai526.com/v1/chat/completions", {
method: "POST",
headers: {
"Authorization": "Bearer YOUR_AI526_API_KEY",
"Content-Type": "application/json"
},
body: JSON.stringify({
"model": "kimi-k3-ad",
"messages": [
{
"role": "user",
"content": "Write a short description of the future."
}
]
}),
});
if (!response.ok) throw new Error(`HTTP ${response.status}`);
console.log(await response.json());
// API URL: use the URL above.
// YOUR_AI526_API_KEY: replace with your own platform key.
// model: use the copied unique ID, or a family ID for the lowest-priced route.
// prompt / messages: replace with your task.Providers for the same API model ID currently share model pricing; variants have independent rates. A selected provider retries only within its own keys, never across variants or providers. Official badges apply to their individual routes.
Model data
Aggregated across providers for the selected variant · UTC
2026-10-30 00:00 — 2026-11-05 07:02 UTC
Request latency
10/3011/05 UTC
Mean end-to-end duration of successful requests, not time to first token.
Request success rate
EarlierLatest
Successful requests divided by total requests, not an SLA.
Output rate
10/3011/05 UTC
Non-estimated output tokens per end-to-end second, shown only from valid logs.
Statistics use UTC. Metrics without valid records display —. Request volume uses the configured public display scale.
Use the API
A family ID chooses the lowest-priced available route; a unique ID selects its own route.
Configure your existing client
Choose a protocol and enter its API URL, your key and this model ID. No client code changes are needed.
- OpenAI · Chat Completions
https://api.ai526.com/v1- Anthropic · Messages
https://api.ai526.com- Model ID
kimi-k3
Base URL: https://api.ai526.com/v1 · /v1/chat/completions
curl https://api.ai526.com/v1/chat/completions \
-H 'Authorization: Bearer YOUR_AI526_API_KEY' \
-H 'Content-Type: application/json' \
-d '{"model":"kimi-k3","messages":[{"role":"user","content":"Write a short description of the future."}]}'
# API URL: use the URL above.
# YOUR_AI526_API_KEY: replace with your own platform key.
# model: use the copied unique ID, or a family ID for the lowest-priced route.
# prompt / messages: replace with your task.import requests
response = requests.post(
"https://api.ai526.com/v1/chat/completions",
headers={
"Authorization": "Bearer YOUR_AI526_API_KEY",
"Content-Type": "application/json"
},
json={
"model": "kimi-k3",
"messages": [
{
"role": "user",
"content": "Write a short description of the future."
}
]
},
)
response.raise_for_status()
print(response.json())
# API URL: use the URL above.
# YOUR_AI526_API_KEY: replace with your own platform key.
# model: use the copied unique ID, or a family ID for the lowest-priced route.
# prompt / messages: replace with your task.const response = await fetch("https://api.ai526.com/v1/chat/completions", {
method: "POST",
headers: {
"Authorization": "Bearer YOUR_AI526_API_KEY",
"Content-Type": "application/json"
},
body: JSON.stringify({
"model": "kimi-k3",
"messages": [
{
"role": "user",
"content": "Write a short description of the future."
}
]
}),
});
if (!response.ok) throw new Error(`HTTP ${response.status}`);
console.log(await response.json());
// API URL: use the URL above.
// YOUR_AI526_API_KEY: replace with your own platform key.
// model: use the copied unique ID, or a family ID for the lowest-priced route.
// prompt / messages: replace with your task.Base URL: https://api.ai526.com · /v1/messages
curl https://api.ai526.com/v1/messages \
-H 'x-api-key: YOUR_AI526_API_KEY' \
-H 'anthropic-version: 2023-06-01' \
-H 'Content-Type: application/json' \
-d '{"model":"kimi-k3","max_tokens":256,"messages":[{"role":"user","content":"Write a short description of the future."}]}'
# API URL: use the URL above.
# YOUR_AI526_API_KEY: replace with your own platform key.
# model: use the copied unique ID, or a family ID for the lowest-priced route.
# prompt / messages: replace with your task.import requests
response = requests.post(
"https://api.ai526.com/v1/messages",
headers={
"x-api-key": "YOUR_AI526_API_KEY",
"anthropic-version": "2023-06-01",
"Content-Type": "application/json"
},
json={
"model": "kimi-k3",
"max_tokens": 256,
"messages": [
{
"role": "user",
"content": "Write a short description of the future."
}
]
},
)
response.raise_for_status()
print(response.json())
# API URL: use the URL above.
# YOUR_AI526_API_KEY: replace with your own platform key.
# model: use the copied unique ID, or a family ID for the lowest-priced route.
# prompt / messages: replace with your task.const response = await fetch("https://api.ai526.com/v1/messages", {
method: "POST",
headers: {
"x-api-key": "YOUR_AI526_API_KEY",
"anthropic-version": "2023-06-01",
"Content-Type": "application/json"
},
body: JSON.stringify({
"model": "kimi-k3",
"max_tokens": 256,
"messages": [
{
"role": "user",
"content": "Write a short description of the future."
}
]
}),
});
if (!response.ok) throw new Error(`HTTP ${response.status}`);
console.log(await response.json());
// API URL: use the URL above.
// YOUR_AI526_API_KEY: replace with your own platform key.
// model: use the copied unique ID, or a family ID for the lowest-priced route.
// prompt / messages: replace with your task.Replace the placeholder key before calling. Media options depend on model support.
Try your own task
Prepare a request, then send it from the dashboard. This page makes no model calls.
Request preview
JSON{
"model": "kimi-k3",
"messages": [
{
"role": "user",
"content": "Write a short description of the future."
}
]
}This is a request preview. Results appear in the dashboard after a request completes.
Frequently asked questions
How should I choose a provider?
Start with automatic routing. To select a public provider, choose its code and copy the updated example. Retries remain within that provider.
How are performance metrics updated?
Metrics update from request records. Latency covers completed successful requests; metrics without valid records display —.
Can I use this model in another client?
The client must support the model’s protocol and a custom API URL. See tool integration guides ↗