moonshot:moonshot/kimi-k3.1-pro-highspeed

Common Name: Kimi K3.1 Pro HighSpeed

Released on Oct 23 12:00 AMSupportedTool InvocationSupportedReasoning

Kimi K3.1 Pro served at 2-3x output speed for latency-sensitive workloads.

Context

1048.6K

Maximum Output

1048.6K

Inputtext, image, video

Outputtext

Collecting…

Availability Trend (24h)

Input$4.40/MTokens

Cached Input$0.22/MTokens

Cache Creation$4.40/MTokens

text.input.cache creation 1h$8.80/MTokens

Output$22.00/MTokens

$2.20/$11.00/M

ctx1.0Mmax1.0Mavail—tps—

InOutCap

Moonshot AI's flagship 3T-class open-weight model with 1M context, strong coding, 3D creation, and financial research.

$1.045/$4.40/M

ctx262Kmax262Kavail—tps—

InOutCap

Moonshot AI's multimodal Kimi K2.6 model for long-horizon coding and agentic work.

$1.045/$4.40/M

ctx262Kmax262Kavail—tps—

InOutCap

Moonshot AI's multimodal reasoning model optimized for coding and agentic software engineering.

$2.09/$8.80/M

ctx262Kmax262Kavail—tps—

InOutCap

Kimi K2.7 Code served at ~180 tokens/s for latency-sensitive coding agents.