@tjrobertson52: Each AI model is the best at something different — here's what I found after testing all three 👀 Which one are you using the most? #ChatGPT #ClaudeAI #Gemini #AITools #AIComparison #Claude #AIForBusiness
Gemini is the laziest model by far - minimal tokens used regardless of model selected
2026-05-22 19:49:02
1
Nads AI Boss Mom 🇯🇲 🇺🇸 :
Yes have multiple subscriptions
2026-06-04 01:01:04
1
Broedkrumme 🇩🇰 :
How much do you walk ?😅
2026-04-29 05:19:10
3
Lorris :
Gemini is getting more lazy lately, they don't even confirm if something is true in search even if you tell him it's false it rely only on current knowledge in pro, I think they try to conserve compute. It wasn't this bad a few month ago. I agree it should be better, I think they limit search on purpose, it would probably be too powerful, war related or security or censorship.
2026-04-29 13:35:09
0
First Name :
I think what you're consistently leaving out but is immensely important is Claude creates its own file archive/documentation/whatever that the other two don't to any meaningful degree and automatically references them when required.
and that set of features alone makes this miles better than any of the other models.
2026-04-29 16:39:27
0
dreamlikepictures :
Have you tested which can best analyze a video link? It seems Gemini is best IMHO. Thanks for all you share.
2026-04-29 21:32:25
1
mon amie :
people are saying perplexity is good for research. have you tried it?
2026-04-29 12:33:26
0
Steve Dave656 :
found Jules recently, was very excited - like codex but definitely not as dependable
2026-04-30 12:30:15
0
janee :
Stop wasting money on 5 different subscriptions. Ryne's ChatAll has the whole collection 😭💯
2026-04-30 21:33:45
1
B R E T T :
Claude Code for planning and Codex for execution of coding
2026-05-01 01:29:50
1
Amori Fati :
Cluade is great but it runs out of limits too quickly even on a max plan. Codex is the new buzz word.
2026-04-29 17:13:01
1
kellygameon :
informative vid!
2026-04-29 14:39:56
0
Eli Peters :
im liking codex... it just seems simpler... less feedback... i did burn thru my 5 hour limit and it struggled with broswer and cli simultaneous use refusing to serve the browser response... ive been opening chatgpt on another free account for simpler shorter answer requests
2026-04-29 09:20:01
1
Cameron Corniuk :
Running these tests have been fun since 5.5 released. I’m finding my workflow shift and change and the results improve. Who knows what it will be next month?!
2026-04-29 10:46:25
1
Lalaocto :
Great summary - thanks for all the informative stuff you put out there. I trust your insight!
2026-04-29 05:01:21
1
aph :
Have you tried Kimi 2.6
2026-05-07 20:17:13
0
Sûni :
❤️❤️❤️
2026-04-29 16:01:16
0
To see more videos from user @tjrobertson52, please go to the Tikwm
homepage.