Meet Kimi Chat (link in last thread) This isn't some side project. It's built by Moonshot AI ,a company founded by ex-Google Brain & Meta AI researchers. They've raised over $1 BILLION from giants like Alibaba & Tencent. They're not here to play games. They're here to undercut

But is it actually good? Kimi's tech is insane: Massive Context: It can handle a 2 MILLION character context window. You can drop entire codebases into it without it breaking a sweat. Elite Coder: It scores 64.2% on SWE-bench and beats Claude-4-Sonnet on tool integration.

Ready to cut your AI bill by 98%? Here's how to get access internationally: - Signup: Kimi asks for a Chinese phone. - Use a service like sms-activate (it costs ~$0.15) to get a temporary number and verify. Bind your email after. - Get API Key: Go to platform.moonsot -

The world of AI is bigger than Silicon Valley. For daily coding, summarization, and document analysis, Kimi Chat delivers competitive performance at a fraction of the cost. Stop burning money. Start building smarter. Bookmark this thread for when your next Claude bill arrives.

Get API Key: Go to http://platform.moonshot.ai Connect to Desktop App (like ChatWise): Base URL: https://api.moonshot.ai/v1 Model: moonshot-v1-32k You've got your cheap claude code in $3 Only.

Google is offering Gemini-2.5-flash-image model tier only for this weekend completely FREE. use and try it- https://x.com/AdityaJiRathore/...

Hey everyone! Quick clarification about the $3 AI alternative: The AI model is GLM-4.5 by http://Z.ai, NOT Kimi Chat. I got my wires crossed while working with multiple AI tools (occupational hazard when you're testing everything 😅). What HASN'T changed: - Same

@AdityaJiRathore Hey Aditya, fantastic post and thank you for the shout-out! You've perfectly captured the incredible value. Just one small clarification: the model in the plan is our own, http://Z.ai's GLM-4.5, not Kimi. But you're 100% right about the performance and the 98%

@AdityaJiRathore It appears there may be some errors in the information presented in the image. The $3 plan is from http://z.ai glm 4.5 model not kimi. Check the subscription link here: https://z.ai/subscribe

@qinkai1028 yes, Thanks for clarifying it, I corrected it.

@AdityaJiRathore This $3 plan comes from the GLM-4.5 model from http://z.ai, not Kimi 😃.

@zRdianjiao Yes you are correct, I really made a big blunder. Working on too many tools and some of them already pile up causes this. Will not repeat this. Thanks for correcting!

@AdityaJiRathore i guess privacy is being priced at $197/mo ?

@bfmcneill Yes, privacy as well as high computing

@AdityaJiRathore I want to cancel ClaudeCode and try, any practical example of the shift to http://z.ai

@meabed Would love to share with you, Although I've found various videos where people complete claude code and this side by side.

@AdityaJiRathore I’m using Trae, it gives 600 messages (to models like Sonet 4) per month for freakin $3…insane.

@soumyattention Pricing is really tempting and attracting

@AdityaJiRathore What about privacy ?

@drago999765 Yes it is chineese, for their own citizen so little to now privacy at all, its a valid concern, that $197 less is for low computing and less privacy.

@AdityaJiRathore Nope, to many hoops to go through, I'll keep cursor and windsurf

@Obama_Biden_13 Me too and in years no tools till now close to them, still we should not confined ourself to some particular tools, its more better to explore other also.

@AdityaJiRathore Great, still it can't compete with claude code.

@CircleProm17860 Correct, but you are getting 70% claude code in just 2% of its price.

@AdityaJiRathore kimi or glm for Coding?

@lprnmns Kimi is better for maximum reliability, speed and especially it is much cost effective but if your workflows heavily relies on tool calling and API reasoning then you can prefer GLM more as it is better for agentic workflows.

@AdityaJiRathore Reality Check: Asked Kimi to code a component library. 5 prompts later, 5USD it loaded up, with several issues. After looping for another 10 USD fixing the bugs I pulled the plug. Claude fixed it in 5min and the project was working. KIMI is cheaper but lacking dev experience

@AdityaJiRathore i love Kimi but part of the sell of Claude code is the amazing TUI they built with it - how are you replacing that? Using the LLM directly on a cli is a big difference

@AdityaJiRathore BS! Z is far inferior than Sonnet 4. That's pure propaganda.

@AdityaJiRathore Could you not also access Kimi in @GroqInc without the Chinese #?

@AdityaJiRathore smells like an android user

@AdityaJiRathore Are the performances good?