kimi-k2-thinking-30%
kimi-k2-thinkingMoonshot Kimi reasoning model interleaving step-by-step thinking with tool calls.
CONTEXT
256K
RELEASED
Nov 2025
INPUT /MTOK
$0.42$0.6
OUTPUT /MTOK
$1.75$2.5
Connect
Call kimi-k2-thinking through the gateway. Native is the model's own wire protocol; Compatible endpoints are translated by the gateway.
POST /v1/responsesNativecurl https://apiarc.ai/v1/responses \
-H "Content-Type: application/json" \
-H "Authorization: Bearer $APIARC_API_KEY" \
-d '{
"model": "kimi-k2-thinking",
"input": "Hello"
}'Providers
Routes offering kimi-k2-thinking through the gateway, with list pricing.
| Provider | Input /M | Output /M | Cache read /M | Latency (30d avg) |
|---|---|---|---|---|
Moonshot Kimi-30% | $0.42$0.6 | $1.75$2.5 | $0.105$0.15 | — |