shadowdrifter
Region: US
Sunday 28 June 2026 04:11:29 GMT
142816
8533
270
354
Music
Download
Comments
Lau :
GLM 5.2 doesn’t compare to opus 4.8 or gpt 5.5 at all, and I say this as someone who has around 4B tokens spent with it and $450+ in pay as you go
2026-06-28 15:41:28
174
BadRally :
used 20 mil tokens on glm 5.2 max for free
2026-06-28 16:15:27
35
.ㅤㅤㅤㅤㅤ :
Told yall
2026-06-30 12:56:02
1
Shadow :
GLM 5.2 is solid bro, but it’s still a hair off from Opus 4.8 imo😭
2026-06-30 10:14:43
3
Dgz :
Sorry, glm 5.2 is 90-97% cheaper than opus 4.8 and it is better
2026-06-29 21:37:23
5
Linux :
Me using a cheap gpt and opus proxy station with a 96% discount on the api models :)
2026-06-29 04:23:28
53
themachinist.sys :
i get unlimited free claude so eh
2026-06-28 09:03:09
13
user870981234 :
Do they have a CLI or Windows Desktop apps like Claude Code and Codex
2026-06-30 10:03:56
1
cantoned :
glm is not opensourse, z.ai not posting weights
2026-06-29 07:03:57
2
Drew Brookins :
Why not run an AI LLM locally on your own computer? I even have Gemma 4 running locally on my phone. Works completely offline.
2026-06-29 17:28:18
1
alxyzz :
I got google ai pro 18 months for $3 and I have on antigravity GPT model and Claude and Gemini
2026-06-29 08:41:02
1
Sophie :
I just checked usage I‘ve used like 70
2026-06-30 11:21:17
1
balanced_7 :
im building a cross model local first knowledge and memory store I called the system Akashic Aurora, its on Github. Basically I want to make it so that any work you do with any agentic Ai gets stored with a narrative spine that links what you did into a searchable narrative that spans multiple domains. it isn't finished but Claude said its been helpful for getting things done in its current state. It even helped me build a just in time context engine that supplies context and learnings from the knowledge store as it works. would anyone be interested in playing with this system and helping me make it better?
2026-06-29 12:45:39
1
lovelacing :
i got 900 million tokens out of deepseek v4 pro for like $15
2026-06-28 23:09:55
8
:
What do you guys use these Ai for, help a brother out mane
2026-06-29 10:26:29
1
Bishop Whitewind :
Sonnet takes 4 hours to do what Opus does in 30 minutes. I can barely imagine the level of pain this takes to do anything complicated.
2026-06-28 14:29:42
30
Farel Pradivi :
9 messages youre joking
2026-06-29 00:04:27
6
Ashen | AI Guy :
yup
2026-06-30 04:38:06
1
momerr1 :
I tested Claude Opus 4.8 and GLM 5.2 Max using the same prompt. Claude was clearly a cut above. If GLM were completely free, it'd be a great option. But with pricing that's almost the same as Claude, I'd definitely choose Claude.
2026-06-28 14:00:41
9
shckzz ♪ :
bro, yk that they were illegally trained on more powerful models, right?
2026-06-29 07:02:36
0
tim :
It’s good but definitely not better
2026-06-29 18:33:29
1
Sam Timmermans :
This is good. Competition means cheaper prices
2026-06-29 16:26:56
3
Lama :
50 million tokens in 1 day
2026-06-29 10:31:34
4
Kalan :
What interface are you using for the open source models
2026-06-29 00:45:55
1
⃟ :
Guys what are you all doing with ai. I just use it for some ai dropshipping but how do you use so many tokens?
2026-06-29 15:58:28
2
To see more videos from user @shadowdrifter4, please go to the Tikwm
homepage.
