Detailed comparison of capabilities, features, and performance.
| Feature | Kimi K2 | Gemini 2.5 Flash Lite Preview |
|---|---|---|
| AI Lab | Moonshot | |
| Context Size | 256,000 tokens | 1,048,576 tokens |
| Max Output Size | 16,384 tokens | 8,192 tokens |
| Frontier Model | No | No |
| Vision Support | No | Yes |
| Description | Kimi K2 (0905) is a state-of-the-art mixture-of-experts (MoE) language model with 32 billion activated parameters and 1 trillion total parameters. Hosted in the USA. | Google's small fast model preview tied to latest version. Features reduced latency and memory requirements while maintaining strong performance on common tasks. |
View detailed specifications and capabilities
View model details →View detailed specifications and capabilities
View model details →Access both Kimi K2 and Gemini 2.5 Flash Lite Preview in a single workspace without managing multiple API keys.
Create your workspace
