Kimi K2.7 Code improves coding scores and token use
Kimi K2.7 Code has been d. Moonshot put “Code” in a K2 model name for the first time, and the shared numbers point to stronger coding . SWE Marathon is up 76.2% compared with K2.6, while token use is down by about 30%.
MLS-Bench Lite rose from 26.7 to 35.1, putting it close to (xhigh) at 35.5. Program-Bench is up 10.4%, and MCP Mark Verified is up 11.4%. The model is also claimed to follow long coding instructions better across full coding tasks.
A high-speed version is planned for beta soon, with 5 to 6 times faster output on the same model.
Key points
- Kimi K2.7 Code is a new coding-focused K2 model from Moonshot.
- SWE Marathon is reported to be 76.2% higher than K2.6.
- Token use is reported to be about 30% lower than K2.6.
- MLS-Bench Lite reached 35.1, close to (xhigh) at 35.5.
- A beta high-speed version is expected to produce output 5 to 6 times faster.