Kimi K2.7 Code released: a coding-focused agent model approaching GPT5.5
Original title:Better Coding with Less Overthinking K2.7 Code takes K2.6's strong base and goes deep Meet @Moonshot...
AI summary
SiliconFlow has launched Moonshot's Kimi K2.7 Code model. Improved on K2.6, it focuses on coding and agent tasks. With 32B activated/1T total parameters, VLM multimodality, and support for interleaved thinking and multi-step tool calls. Compared with K2.6, reasoning token usage is reduced by 30%, cutting overthinking; instruction following and completion rates on long-horizon coding tasks improved. Performance approaches GPT5.5 and Opus 4.8. Pricing: cached input 0.19/input 0.94/output 4.00 per million tokens.
Key Takeaways
- SiliconFlow has launched Moonshot's Kimi K2.
- 7 Code model.
- Improved on K2.
- 6,it focuses on coding and agent tasks.
- With 32B activated/1T total parameters,VLM multimodality,and support for interleaved thinking and…
Quick Facts
- Published
- Updated
- Category
- Model Releases
- AI score
- 65
K2.7 Code 在编码上逼近 GPT-5.5 和 Opus 4.8,同时推理 token 用量降低 30%,对追求前沿编码能力又在意成本的技术团队是个实用选择,硅基流动上的定价也给了明确预期。
Related entities
Source & copyright
This article is based on curated titles and AI summaries from the AIHOT public API, with attribution and links back. Original copyright remains with the original publishers and authors; this site does not mirror full texts.
How to Cite
When citing this article, keep the title, date, and permalink, and credit Orange News as the source.
Orange News: Kimi K2.7 Code released: a coding-focused agent model approaching GPT5.5. June 16, 2026. https://new.nat6.net/en/news/cmqgq9igu022aslic2gclug83/