Kimi K2 0905 is the September update to Kimi K2 0711, a large-scale Mixture-of-Experts (MoE) model developed by Moonshot AI. It features 1 trillion total parameters with 32B active per forward pass and extends long-context inference from 128K to 256K tokens.
This release enhances agentic coding with improved accuracy and better generalization across scaffolds, while also boosting frontend development with more refined and functional outputs for web, 3D, and related applications. Optimized for agentic capabilities—spanning advanced tool use, reasoning, and code synthesis—Kimi K2 continues to excel across benchmarks in coding (LiveCodeBench, SWE-bench), reasoning (ZebraLogic, GPQA), and tool use (Tau2, AceBench). Training is powered by a novel stack that incorporates the MuonClip optimizer for stable, large-scale MoE performance.
| Creator | Moonshot AI |
| Release Date | September, 2025 |
| License | Modified MIT License |
| Context Window | 262,144 |
| Image Input Support | No |
| Open Source (Weights) | Yes |
| Parameters | 1000B, 32.0B active at inference time |
| Model Weights | Click here |
Performance Benchmarks
| Benchmark | Metric | K2-Instruct-0905 | K2-Instruct-0711 | Qwen3-Coder-480B-A35B-Instruct | GLM-4.5 | DeepSeek-V3.1 | Claude-Sonnet-4 | Claude-Opus-4 |
|---|---|---|---|---|---|---|---|---|
| SWE-Bench verified | ACC | 69.2 ± 0.63 | 65.8 | 69.6* | 64.2* | 66.0* | 72.7* | 72.5* |
| SWE-Bench Multilingual | ACC | 55.9 ± 0.72 | 47.3 | 54.7* | 52.7 | 54.5* | 53.3* | – |
| Multi-SWE-Bench | ACC | 33.5 ± 0.28 | 31.3 | 32.7 | 31.7 | 29.0 | 35.7 | – |
| Terminal-Bench | ACC | 44.5 ± 2.03 | 37.5 | 37.5* | 39.9* | 31.3* | 36.4* | 43.2* |
| SWE-Dev | ACC | 66.6 ± 0.72 | 61.9 | 64.7 | 63.2 | 53.3 | 67.1 | – |
