Skip to content
Models/kimi/Kimi K3

Kimi K3

DashScope·kimi
kimi-k3Text·

Moonshot AI flagship open MoE model with 2.8T parameters, 1M context, native vision understanding, and always-on deep reasoning.

Context Window
1.0M
Input Price /M in
$3.08$2.77/M in
Output Price /M out
$15.38$13.85/M out
Cached Input Price /M
$0.28/M
Max Completion
131K
Input Modalities
text, image
Output Modalities
text
reasoningFunction callingChatVisionJSONStreamingNew-10%GeneralProgrammingLongContext

Description

Moonshot AI flagship open MoE model with 2.8T parameters, 1M context, native vision understanding, and always-on deep reasoning.

Available Providers

AllToken can route requests to the providers below based on route priority and policy.

ProviderContextInputOutputCached / MLatencyThroughput

Best For

Moonshot AI flagship open MoE model with 2.8T parameters, 1M context, native vision understanding, and always-on deep reasoning.

How To Use This Model

Use the exact model ID shown below. This is the safest way to avoid call failures, variant mismatches, or incorrect route assumptions.

curl https://api.alltoken.ai/v1/chat/completions \
  -H "Authorization: Bearer sk-your-key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "kimi-k3",
    "messages": [
      {"role": "user", "content": "Hello!"}
    ]
  }'
Supported Parameters
temperaturetop_pmax_tokenstoolsresponse_formatreasoning_effortweb_search
API Key Setup
Smart Routing

Let the platform choose the best provider path automatically.

Default Model

If a request does not specify a model, default the key to kimi-k3.

Forced Model

Always override incoming requests and lock the key to kimi-k3.