LLM Reference
NVIDIA NIM

Kimi K2.5 on NVIDIA NIM

Kimi · Moonshot AI

Serverless

Pricing

TypePrice (per 1M)
Input tokensFree
Output tokensFree

Capabilities

VisionMultimodalReasoningFunction CallingTool UseJSON ModeCode Execution

About Kimi K2.5

Moonshot AI's Mixture-of-Experts large language model with 1 trillion total parameters across 384 experts. Strong performance on coding tasks (73.2 SWE-Bench Verified). Supports function calling via tool calls. JSON mode support unconfirmed as of March 2026. Available via platform.moonshot.cn API.

Get Started

Model Specs

Released2026-03-15
Parameters1T (MoE, 384 experts)
Context256K
ArchitectureMixture of Experts

Related Models on NVIDIA NIM