Benchmarks of rumored Mythos level model from Zhipu AI

binyu2 pts0 comments

Ananay on X: "Ox Alpha is 100% a GLM model by Zhipu AI, and it looks like its *almost* mythos class from very early results.

It's very likely going to be called GLM-6 and it's absolutely mogging every frontier model in SWE and Cyber benchmarks (DeepSWE screenshot below)

🧵 (1/n)" / X<br>Post

Log inSign up

Post

Ananay on X: "Ox Alpha is 100% a GLM model by Zhipu AI, and it looks like its *almost* mythos class from very early results.

It's very likely going to be called GLM-6 and it's absolutely mogging every frontier model in SWE and Cyber benchmarks (DeepSWE screenshot below)

🧵 (1/n)"

Ananay

@ananayarora

Ox Alpha is 100% a GLM model by Zhipu AI, and it looks like its *almost* mythos class from very early results.

It's very likely going to be called GLM-6 and it's absolutely mogging every frontier model in SWE and Cyber benchmarks (DeepSWE screenshot below)

🧵 (1/n)

span:not(:empty)~span:not(:empty)]:before:content-['·'] [&>span:not(:empty)~span:not(:empty)]:before:px-1 [&>span:not(:empty)~span:not(:empty)]:before:shrink-0">10:06 AM · Aug 21, 202660.6KViews

26<br>22<br>578<br>192

span:not(:empty)~span:not(:empty)]:before:content-['·'] [&>span:not(:empty)~span:not(:empty)]:before:px-1 [&>span:not(:empty)~span:not(:empty)]:before:shrink-0 min-w-0 overflow-hidden">Ananay

@ananayarora

8h

First off, Zhipu AI codenames its models with names of Chinese Zodiac signs followed by "Alpha". They did this in Feb 2026 with GLM 5 codenaming it "Pony Alpha" and releasing it later as GLM-5. The same was released on OpenRouter as stealth/pony-alpha. Hence, Ox Alpha. (2/n)

52<br>4.9K

span:not(:empty)~span:not(:empty)]:before:content-['·'] [&>span:not(:empty)~span:not(:empty)]:before:px-1 [&>span:not(:empty)~span:not(:empty)]:before:shrink-0 min-w-0 overflow-hidden">Ananay

@ananayarora

8h

Ox Alpha's tokenizer is one for one identical to GLM's. No other lab's tokenizer gets past 4/11 when it comes to tokenizer similarity.<br>claude.ai/code/artifact/…<br>(h/t: @sushsrinivasan)

(3/n)

58<br>5K

span:not(:empty)~span:not(:empty)]:before:content-['·'] [&>span:not(:empty)~span:not(:empty)]:before:px-1 [&>span:not(:empty)~span:not(:empty)]:before:shrink-0 min-w-0 overflow-hidden">Ananay

@ananayarora

8h

Ox Alpha is free right now, unlimited usage and no guardrails on OpenRouter and OpenCode. If this model drops as an open source model and really turns out to be as good as mythos on all coding & cyber benchmarks it might just be the final nail in the coffin for US AI labs.

71<br>3.7K

span:not(:empty)~span:not(:empty)]:before:content-['·'] [&>span:not(:empty)~span:not(:empty)]:before:px-1 [&>span:not(:empty)~span:not(:empty)]:before:shrink-0 min-w-0 overflow-hidden">Petri Kuittinen

@KuittinenPetri

3h

I have changed my mind after probing stealth/0x-alpha that it is not Chinese model. It is based on GLM-5.x series, but post-trained by a Western company, mostly likely XAI, as many of the reasoning sound like Grok. For China sensitive topics (except Taiwan) it is generally giving Show more

19<br>1.3K

Log in or sign up for X<br>See what’s happening and join the conversation<br>Continue with phoneContinue with AppleContinue with Google<br>or<br>Log in with username or email

Relevant people

Ananay@ananayaroraFollow<br>Building @OpenHackAI, previously @Apple.

Trending now

span empty before model alpha ananay

Related Articles