@cognition: Try Kimi K2.7 and GLM 5.2 for free in Devin Desktop and CLI
Summary
Devin Desktop now supports Kimi K2.7 and GLM 5.2 models, offering free trials until July 5 for Pro/Max/Teams users.
View Cached Full Text
Cached at: 06/25/26, 01:19 PM
Try Kimi K2.7 and GLM 5.2 for free in Devin Desktop and CLI
Devin Desktop (@devindesktop): Kimi K2.7 Code and GLM 5.2 are available in Devin Desktop and CLI
Both perform strongly on FrontierCode Extended, our benchmark for real-world engineering tasks: GLM 5.2: 43.0% Kimi K2.7 Code: 39.5%
Pro/Max/Teams users can try both models for free until July 5
Similar Articles
Has anyone noticed that the behavior of the Kimi model has changed?
A user observes that the Kimi K2.6 model's chain-of-thought has become shorter and more concise, improving coding performance in Kimi Code, and expresses hope for continued open-source competition with upcoming GLM 5.2 and Fable 5.
@atomic_chat_hq: New @Zai_org GLM-5.2 beats Kimi K2.7 Code on physics contest! We gave both models the same three prompts and asked them…
Z.ai releases GLM-5.2, an open-weights AI model with improved coding and agentic performance, demonstrated by beating Kimi K2.7 Code on a physics simulation benchmark across three tasks.
@LexnLin: why tf is GLM 5.2 and Kimi 2.7 unlimited in a Devin sub rn? literally a gold mine
The tweet highlights that GLM 5.2 and Kimi 2.7 are available without limits in a Devin subscription, describing it as a gold mine.
UPDATE: "Gentle Coding" is mathematically proven. 1,500+ test runs show major gain for Kimi K2.6 and even more for GLM-5.1! GPT 5.4/5.5 and Claude Sonnet 3.5/Opus 4.6 also better, with ZERO REGRESSION ACROSS THE BOARD.
The 'Gentle Coding' technique is empirically validated across 1,500+ tests, showing significant improvements (zero regression) for multiple models including Kimi K2.6, GLM-5.1, GPT 5.4/5.5, and Claude Sonnet 3.5/Opus 4.6 by reducing looping and hallucinations.
Kimi K2.6
Kimi K2.6 is released as an open-source model that achieves state-of-the-art performance on long-horizon coding and agent swarm benchmarks.