k3
Moonshot
Multimodal reasoning model for complex coding, knowledge work, and long-running agent tasks. Works across large codebases, calls tools, and debugs by leveraging images, logs, and test outputs to continuously improve results.
- Context
- 1M
- Input / 1M tokens
- ¥21.6
- Output / 1M tokens
- ¥108
- Cache read / 1M
- ¥2.16
- Cache write / 1M
- -