Kimi K2 Thinking (Mumbai)

Foundation model available in Mumbai. Direct inference without cross-region routing.

Provider: All Moonshot models | AWS Bedrock

Inference regions: ap-south-1

API Endpoint

https://bedrock-runtime.ap-south-1.amazonaws.com

Supported Parameters

ParameterTypeDescription
max_tokensintegerMaximum number of tokens to generate in the response. (≥1)
temperaturefloatControls randomness. Lower values make output more deterministic. (0–1) Default: 1.
top_pfloatNucleus sampling threshold. Considers tokens with cumulative probability up to this value. (0–1) Default: 1.
stop_sequencesstringUp to 4 sequences where the model will stop generating.
top_kintegerOnly sample from the top K most likely tokens at each step. (0–500) Default: 250.

Feature Guides

On-Demand Inference

Default mode. Pay per token with no upfront commitment.

Documentation