Marcin AI
Region: US
Thursday 28 May 2026 19:16:43 GMT
46233
1316
43
212
Music
Download
Comments
therealargonate :
my first task with 4.8 failed and I had to have Codex fix it.
2026-05-28 19:38:36
40
Quality_Expert_Writers :
should I resubscribe Claude now?. I mainly use codex
2026-05-28 20:48:16
1
Ramiel :
I don't think Claude will ever be as good as it was, the usage limits are are a problem I mean you ask it run an audit it runs out halfway, I asked codex the same thing it fixed and found errors unrelated to what I asked.
2026-05-29 01:04:41
16
fokketje :
Bros caption is ai
2026-05-29 07:51:27
4
kalspersvi :
it's gonna get nerfed for some reason next week
2026-05-28 22:11:25
21
Ovea :
4.7 was tough it made you question your idea it chalenges you I don’t think it was a fumble at all in my opinion
2026-05-29 00:44:13
10
Matthew McCall-Stillman :
Are you using Hermes?
2026-06-02 01:59:38
1
John McLaughlin :
Hasn’t been the same since 4.6. It refuses to commit to and complete a task. Gets distracted too easily.
2026-05-29 03:57:19
4
Zakch kbir :
how ots 3x cheaper when price per 1m token the same
2026-05-29 00:00:40
4
Noah🦦 :
Why do they compare it to 3.1 pro. 3.5 flash is better and 3.5 pro comes out next month anyways
2026-05-29 13:19:04
0
Ngai He Hakka Nhin :
Biggest turn off for me was the ridiculous rate limit,.which was why I stopped using it. Regarding sub agents swamp, it only use 16. Kimi uses a.bigger swamp.
2026-05-31 14:05:59
0
Megan Writes :
Yesss thanks for the breakdown!
2026-05-30 01:51:06
1
Thorne-Curtis Joe :
Is it as good as 4.6 before the nerf? are we just being trained to accept mediocrity and assume that each model is "better" because their benchmarks say so?
2026-05-28 23:37:05
3
Danielle Unpacks AI :
Using it all morning
2026-05-28 20:06:17
1
M. Salim Zaki :
I haven’t used it yet but I’m still giving Claude the website design tasks not something critical
2026-05-28 22:05:52
0
noah :
nice caption
2026-05-29 13:54:22
0
bukan babumu :
its suck bro.. wth
2026-05-29 02:58:48
3
JohannesDB🇩🇰 :
At this point: it the model affect your workflow that much - your harness and workflow is wrong
2026-05-29 15:36:59
0
Anbu_Yeagerist :
I like it
2026-05-30 11:33:53
0
Sebastian Wolff :
Finally. I actually switched back to 4.6, that’s how unhappy I was with 4.7
2026-05-29 05:48:42
0
sando_ai :
I have been testing 4.8 and it is a lot smoother and it is not killing the tokens as much, but the dynamic workflows burns tokens quick, and I tested against the codex app and codex burns a lot more tokens, good video mate
2026-05-29 12:44:36
0
Matrix guy :
Codex is still better
2026-05-29 10:26:18
0
Mysterious :
compare with deepseek,
2026-05-28 20:06:42
1
Super Crafting :
Trust me bro benchmark
2026-05-29 06:38:22
1
AI & Automations :
good luck 👌
2026-05-28 19:26:19
0
To see more videos from user @marcinteodoru, please go to the Tikwm
homepage.
