@berryxia: OMG, I'm totally blown away by Kimi 3! Incredible. It feels like it crushes Kimi 2.7! Let's first look at the test with the same prompt on the previous 2.7 model.
Summary
After testing Kimi 3, the user found its performance far exceeds the previous generation Kimi 2.7, calling it a 'crushing' victory, and mentioned that a demo is already available on the official website.
View Cached Full Text
Cached at: 07/17/26, 02:26 AM
Holy shit, Kimi 3 is seriously impressive! It’s damn good.
Feels like it completely blows Kimi 2.7 out of the water!
Let’s first take a look at the test results from the previous generation 2.7 model with the same prompt. https://t.co/zfokEAS2kv
Berryxia.AI (@berryxia): Guys, Kimi 3 has quietly been released on the official site! After watching a bunch of demos today, I have to say it’s really stepped up this time~
Go try it out now! Link below 👇🏻
Similar Articles
@berryxia: Guys, feels like 3D modeling CAD is 'dead'? Kimi 3 directly dominates the top models! The quality and interaction are seamless, you don't even need complex prompts. No nonsense, just watch the video.
The author praises the Kimi 3 model's performance in 3D modeling CAD, stating that its texture and interaction surpass the previous generation Kimi 2.7, and includes a video demonstration.
@jakevin7: Using Kimi K3, Maka outperforms official KimiCode by 20%. Same model, different harness — how big can the gap be? http://github.com/maka-agent/maka-agent… We ran Kimi K3 through…
On the Kimi K3 model, the open-source agent framework Maka achieves a 10% higher overall pass rate on Terminal-Bench 2.1 compared to the official Kimi Code CLI, and 20% higher on hard tasks. Through optimizations like context-budget pruning, streamlined tool surface, and concise system prompts, significant performance gains are realized. The full report and harness are open-sourced.
@berryxia: Looks like Kimi has been quietly making big moves. After watching these demo videos, my back didn't get cold — instead, I got pumped and felt truly impressed! Kimi today directly upgraded AI agents from "just chatting" to "operating browsers like humans." They released the Web Bridge browser extension.
Kimi has released the Web Bridge browser extension, enabling AI agents to operate browsers like humans, performing tasks such as searching, scrolling, and clicking, while supporting multiple AI tools.
@__oQuery: Everyone says K3 is easy to use, so I opened one and tried the latest Kimi Code. But I found that Kimi's development team might really not have a good feel for design. As soon as I started using it, I noticed many frustrating issues: 1. The ask user question in figure one…
User posts criticism of Kimi Code's interaction design, pointing out issues such as the unreasonable submission method of the ask user question tool and the state bar shaking due to container height changes during command selection, hoping the Kimi team will optimize.
@seclink: Compared to Mimo, Kimi is slower and more expensive ...
Kimi released the K2.7 Code model and its high-speed version, and announced API pricing. Compared to rival Mimo, it is more expensive and slower.