🔥 gpt-6-astra is Now Live – First Release! Try Now
Sign in
Product
ModelsToolsPlaygroundCompute
Resource
BlogsRankingDocumentationUse CasesStatus
CLIPricing
Language
Back to ModelsGet startedView API

kimi-k3 API

kimi-k3

Kimi K3 is Moonshot AI's flagship model for long-horizon coding and knowledge work. Moonshot documents a 1M-token context, native visual understanding, and open-weight availability; Flatkey exposes compatible chat and messages routes with the catalog rates below.

Chat and codingLong contextTool workflows
Input /M
Reference price: $3
$2.4-20%
Cache /M
Reference price: $0.3
$0.24-20%
Output /M
Reference price: $15
$12-20%
Provider
Moonshot AI

Performance

Kimi K3 API performance and availability

Live Flatkey request telemetry appears here when enough Kimi K3 traffic is available; no benchmark ranking is inferred.

Avg. provider uptime
93.7%
last 30 days
Latency
9.03s
last 30 days
Requests
20.4K
30-day window
Successful inference trend

Activity

Kimi K3 API usage and request activity

This chart uses live Flatkey request data for Kimi K3 and stays unreported until enough traffic is collected.

Requests
20.4K
30-day window
Latency
9.03s
last 30 days
Uptime
93.7%
last 30 days
Successful inference trendActivity

Pricing

Kimi K3 API pricing for long-context requests

Prices below are calculated from Flatkey pricing data for this model and the visible groups currently returned by our pricing API.

Flatkey price

Input /M

Live catalog model
$2.4
Input /M
$2.4$3
Cache /M
$0.24$0.3
Output /M
$12$15
View API
Shared balance

Add credits

Add credits

Use the same Flatkey balance and API key across image, video, audio, and text models.

$10Open wallet$20Open wallet$50Open wallet
Model catalog
Model Type
Text
API
/v1/chat/completions
Billing basis
1M tokens
Context
1,048,576 tokens

Kimi K3 capabilities for coding, files, and research

These cards describe practical workflows; upstream facts and Flatkey route fields remain separated below.

Long-context knowledge work

Work across long documents, codebases, and research notes while keeping the relevant material in one workflow.

Vision and file-aware analysis

Use visual understanding upstream and the file inputs listed for the hosted route when a task needs richer context.

Coding and research assistance

Draft, inspect, explain, and transform technical material for engineering and knowledge teams.

Flexible client integration

Bring the model into compatible chat or messages clients while keeping the same model identity and account controls.

Compare

Kimi K3 hosted API vs open-weight and local options

Moonshot documents open-weight availability upstream; this page describes Flatkey's hosted API route and does not promise native local execution inside Flatkey.

CapabilityMoonshot open-weight / local optionVerified Flatkey hosted API
Hosted endpointUnknown/v1/chat/completions and /v1/messages
ContextUnknown1,048,576 tokens
Input modalityNative vision plus documented upstream inputsText/file fields in Flatkey catalog
Downloadable weightsDocumented by Moonshot upstreamFlatkey local delivery not verified
Free accessUnknownNot promised; paid token rates apply
Upstream vs FlatkeyMoonshot open-weight release and local use are upstream optionsFlatkey routes and bills hosted API access

Why use Flatkey for Kimi K3?

A long-context Kimi K3 workflow for coding and research

Make the two compatible routes, file input, and token dimensions visible before you build an agent or knowledge workflow.

Two client paths

Choose /v1/chat/completions for an OpenAI-shaped client or /v1/messages for an Anthropic-shaped client.

Document-first use cases

Use the verified file modality and 1,048,576-token context for long documents, codebases, and research notes.

Hosted boundary is clear

Moonshot's open-weight release is an upstream option; this page only promises Flatkey hosted API access, not native local hardware support.

Predictable token accounting

See input, output, and cache rates in the pricing block before scaling a workflow.

API

Use the Kimi K3 API with OpenAI or Anthropic clients

Choose the compatible route documented by your client and set model to kimi-k3; Moonshot upstream access and Flatkey hosted routing are separate services.

kimi-k3
POST /v1/chat/completions
POST /v1/messages
Compatible routes do not prove local or open-source support.
curl https://router.flatkey.ai/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer sk-..." \
  -d '{
  "model": "kimi-k3",
  "messages": [
    {
      "role": "user",
      "content": "Extract the action items from this document and return owners, deadlines, and open questions."
    }
  ]
}'
Docs

FAQ

Kimi K3 API
pricing and local-use questions

Answers about Kimi K3 API routes, model ID, long context, file input, pricing, hosted access, and local-deployment boundaries.

What is Kimi K3?
Kimi K3 is Moonshot AI's long-context model with a documented 1,048,576-token context and native visual understanding. Flatkey exposes a hosted compatible API route.
How do I use Kimi K3?
Create a Flatkey API key, choose /v1/chat/completions or /v1/messages, and set model to kimi-k3.
What is the Kimi K3 API model ID?
The verified model ID is kimi-k3.
How much does Kimi K3 cost?
Use the pricing section above for current Flatkey prices from our pricing API.
Is Kimi K3 free?
The current catalog shows paid token rates; do not promise free access.
Is Kimi K3 open source or available locally?
Moonshot documents an open-weight release that can be downloaded upstream. Flatkey's page provides hosted API access; native Flatkey local deployment and hardware requirements are not promised.
Who makes Kimi K3?
The catalog vendor is Moonshot AI; Flatkey provides routing and billing.