All Models
Kimi K2 Instruct
Kimi K2 Instruct is a large-scale Mixture-of-Experts (MoE) language model developed by Moonshot AI, featuring 1 trillion total parameters with 32 billion active per forward pass. It is optimized for agentic capabilities, including advanced tool use, reasoning, and code synthesis. Kimi K2 excels across a broad range of benchmarks, particularly in coding (LiveCodeBench, SWE-bench), reasoning (ZebraLogic, GPQA), and tool-use (Tau2, AceBench) tasks. It supports long-context inference up to 128K tokens and is designed with a novel training stack that includes the MuonClip optimizer for stable large-scale MoE training.
Available Providers (13)
| Provider | Model ID | Input Cost | Output Cost | Context | Max Output | Docs |
|---|---|---|---|---|---|---|
 Nvidia | moonshotai/kimi-k2-instruct | $0.00/MTok | $0.00/MTok | 128K | 8.2K | |
 NanoGPT | moonshotai/kimi-k2-instruct | $0.10/MTok | $2.00/MTok | 256K | 8.2K | |
 IO.NET | moonshotai/Kimi-K2-Instruct-0905 | $0.39/MTok | $1.90/MTok | 32.8K | 4.1K | |
 Nebius Token Factory | moonshotai/Kimi-K2-Instruct | $0.50/MTok | $2.40/MTok | 200K | 8.2K | |
 Vultr | Kimi-K2.5 | $0.55/MTok | $2.75/MTok | 261K | 32.8K | |
 Cortecs | kimi-k2-instruct | $0.55/MTok | $2.65/MTok | 131K | 131K | |
 NovitaAI | moonshotai/kimi-k2-instruct | $0.57/MTok | $2.30/MTok | 131.1K | 131.1K | |
 Jiekou.AI | moonshotai/kimi-k2-instruct | $0.57/MTok | $2.30/MTok | 131.1K | 131.1K | |
 Groq | moonshotai/kimi-k2-instruct | $1.00/MTok | $3.00/MTok | 131.1K | 16.4K | |
 Fireworks AI | accounts/fireworks/models/kimi-k2-instruct | $1.00/MTok | $3.00/MTok | 128K | 16.4K | |
 Together AI | moonshotai/Kimi-K2-Instruct | $1.00/MTok | $3.00/MTok | 131.1K | 131.1K | |
 Vercel AI Gateway | moonshotai/kimi-k2 | $1.00/MTok | $3.00/MTok | 131.1K | 16.4K | |
 Hugging Face | moonshotai/Kimi-K2-Instruct | $1.00/MTok | $3.00/MTok | 131.1K | 16.4K |
Capabilities
Reasoning
Tool Calling
Attachments
Open Weights
Structured Output