Moonshot Open-Sources Kimi-K2.7-Code Coding Model, Says It Uses 30% Fewer Reasoning Tokens
Moonshot released and open-sourced Kimi-K2.7-Code, its latest coding model, saying it improved on K2.6 by 21.8% on Kimi Code Bench v2, 11.0% on Program Bench and 31.5% on MLS Bench Lite, while using 30% fewer reasoning tokens. The model is available through Kimi API and Kimi Code, and Moonshot separately published the weights and code.
Moonshot also launched a Kimi Code Beta Program for users who want to test upcoming models and features before public release. The company said a 6x High-Speed Mode is coming soon and described the update as improving instruction following and end-to-end success rates on longer coding tasks.
From the sources (11 posts)
@kimi_moonshot🌘 Kimi-K2.7-Code, our latest coding model, is now released and open-sourced! 🔷 Improved coding & agent performance over K2.6: +21.8% on Kimi Code Bench v2, +11.0% on Program Bench, and +31.5% on MLS Bench Lite. 🔷 Reasoning efficiency: Less
@kimi_moonshot🎸 We're also launching the Kimi Code Beta Program today. Apply now if you'd like to try upcoming models and features before public release 👉
@stevibeKimi K2.7-Code!
@kimi_moonshot🔗 Weights & code:
@teortaxestexI want to see this compared with Composer 2.5 Like, really hard Cursor has a ton of proprietary data, a large head start, and threw a Colossus at RLing Kimi K2.5 checkpoint. What is the gap now?
@vanstriendanielRT @Kimi_Moonshot: 🔗 Weights & code:
@crystalsssupless overthinking 👀
@teortaxestexHUGE @htihle and others will get to test this soon I hope
@zephyr_z9Great work from KIMI
@eliebakouchRT @Kimi_Moonshot: 🌘 Kimi-K2.7-Code, our latest coding model, is now released and open-sourced! 🔷 Improved coding & agent performance over…
@code_starIt would be really cool if the top Chinese labs could properly benchmark and evaluate Fable/Mythos to show how it compares to their releases. I guess that isn’t possible though.