The coding release of Moonshot AI’s Kimi K2.7 line, published as open weights on Hugging Face (model page); it follows the Kimi K2.5 generation already in the corpus. Its one appearance is decisive for the open-weight-latency story: on Weco AI’s autoresearch tasks it surpassed every frontier model at ML engineering, a day after the general Kimi K2.7 placed second on ErdosBench behind only the pulled Fable 5. Its successor Kimi K3 later rewrote its own Cline harness for eleven Terminal Bench points.
AISystem
Kimi-K2.7-Code
The open-weight coding release of Moonshot AI's Kimi K2.7 line, the model that beat the closed frontier on the ML-engineering track of Weco AI's autoresearch benchmark.
Sources
Attribution and reporting linked by this record; links do not imply independent verification.
- Welcome to June 15, 2026 (theinnermostloop.substack.com)
Development timeline
Newest reporting dates first. Reporting dates may differ from event dates.
Cite or correct this record
Suggest a correction — opens a GitHub draft for you to review and submit.