Kimi K3
meldport/kimi-k3Kimi K3 through the Meldport API for production AI workloads.
MODEL OVERVIEW
About Moonshot AI Kimi K3
Kimi K3 is a verified model available through Meldport. It uses a stable public alias, reviewed capability metadata, and an explicit customer price. Meldport selects a healthy provider for the alias and records the observed cost boundary for reconciliation and usage accounting.
GOOD FIT FOR
- Text generation, extraction, and automation
- Image-aware prompts and visual analysis
- Tool calling and agent workflows
BEFORE YOU SHIP
- Check capabilities, context limit, and price against your workload.
- Test tool call parsing and error handling before production.
- Verify image input size and format limits for your use case.
TECHNICAL SPECIFICATIONS
What this model supports
| Model ID | meldport/kimi-k3 |
|---|---|
| Publisher | Moonshot AI |
| Capabilities | Text, Vision, Cache, Tools |
| Context window | 1M |
| Lifecycle | active |
| Verified | Nov 1, 2026 |
CURRENT RATES
Pricing
| Metric | Rate | Unit |
|---|---|---|
| Input | $0.9 | per million tokens |
| Output | $4.5 | per million tokens |
| Provider-reported cached input | $0.09 | per million tokens |
| Minimum charge | $0.00001 | per request |
Provider-reported cached input tokens use the displayed rate. Every request is recorded in the wallet ledger.
QUICKSTART
Call this model
The same endpoint works from any language.
Send POST /v1/chat/completions with an API key that has inference:invoke. Use meldport/kimi-k3 in the model field. Supply your conversation in messages and set stream: true for streamed output. The published context limit is 1M tokens.
This model publishes tool calling. Send function definitions in tools and return tool results in the conversation. Vision inputs are listed for this model. Request fields and output limits · Streaming format · Errors and retry rules
curl https://api.meldport.com/v1/chat/completions \
-H "Authorization: Bearer $MELDPORT_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "meldport/kimi-k3",
"messages": [{"role": "user", "content": "Hello"}]
}'import os
import requests
response = requests.post(
"https://api.meldport.com/v1/chat/completions",
headers={
"Authorization": f"Bearer {os.environ['MELDPORT_API_KEY']}",
},
json={
"model": "meldport/kimi-k3",
"messages": [{"role": "user", "content": "Hello"}],
},
)
print(response.json())const response = await fetch(
"https://api.meldport.com/v1/chat/completions",
{
method: "POST",
headers: {
Authorization: "Bearer " + process.env.MELDPORT_API_KEY,
"Content-Type": "application/json",
},
body: JSON.stringify({
model: "meldport/kimi-k3",
messages: [{ role: "user", content: "Hello" }],
}),
},
);
console.log(await response.json());COMMON QUESTIONS
Frequently asked questions
What is the API model ID for Kimi K3?
Use meldport/kimi-k3 in the model field of your API request.
Which capabilities does Kimi K3 support?
This model currently publishes Text, Vision, Cache, Tools support through Meldport.
How is Kimi K3 priced?
Requests are billed per token. Input costs $0.9 per million tokens and output costs $4.5 per million tokens. Provider-reported cached input tokens cost $0.09 per million tokens; Meldport response caching is bypassed. A minimum charge of $0.00001 applies per request. Charges are deducted from your wallet balance in real time.
Which endpoint runs Kimi K3?
Send a request to POST https://api.meldport.com/v1/chat/completions with your Meldport API key and use meldport/kimi-k3 in the model field.
What context window does Kimi K3 support?
The published context limit is 1M tokens.