Kimi K2
Kimi K2 is Moonshot AI’s original non-thinking agentic MoE language model, with 1T total and 32B activated parameters. MuonClip-based training and large-scale agent post-training target coding, instruction following, and tool use. The July 11 release offers base and instruction checkpoints with a 128K context. The September 5 Instruct update improves coding and extends context to 256K, grouped here as a weight variant. The separately released K2-Thinking reasoning model has its own page. Public weights use Modified MIT terms; tools require an external execution harness.
2025-07-11
1T total, 32B activated
Mixture of Experts (MoE)
Modified MIT
Specifications
- Parameters
- 1T total, 32B activated
- Architecture
- Mixture of Experts (MoE)
- License
- Modified MIT
- Context Window
- 128,000 tokens
- Type
- text
- Modalities
- text
Advanced Specifications
- Model Family
- Kimi
- API Access
- Available
- Chat Interface
- Available
- Multilingual Support
- Yes
- Variants
- Kimi-K2-BaseKimi-K2-InstructKimi-K2-Instruct-0905 (256K context, 2025-09-05)
Capabilities & Limitations
- Capabilities
- reasoningcodemathagentictool-usemultilinguallong-context
- Known Limitations
- excessive token generation on hard reasoningperformance decline with tool use on certain tasksincomplete tool calls with unclear definitions
- Notable Use Cases
- agentic applicationscoding assistantdata analysisweb developmentresearch tasks
- Function Calling Support
- Yes
- Tool Use Support
- Yes