Luke
Region: US
Saturday 15 August 2026 23:02:38 GMT
5520
156
29
8
Music
Download
Comments
user82583518251 :
Claude Numbus is suppose to have double the intellectual power of fable
2026-08-15 23:55:58
1
Eabusham2 :
Claude only does 50% what I say and I have to keep telling it to check itself I even made a skill called /donecheck for after and /checklist when making and it’s kinda works but still not 100%
2026-08-15 23:41:42
2
HIM :
Do you recommend spending money DeepSeek fast or pro v4 (my friend needs help in his Roblox game he use Claude i ain’t that rich)
2026-08-16 00:13:52
1
harley :
terminal 2.1 is an outdated bench, u have to compare all the benches bruh
2026-08-16 14:33:48
0
alovelydayyyyyy :
True... sucking up my tokens.
2026-08-18 05:50:40
1
Jakob Schulz :
Your point might be good but comparing token prices is not showing anything. Tokens are not a standard unit, a response doesn't use a fixed amount of tokens, even input token amount can differ for the same input. Please use something like price per task (maybe in the bench you are already referencing.)
2026-08-17 00:38:05
0
drgt :
Lol terminal bench 2.1 its so trash
2026-08-16 13:20:08
0
shampeh.ai :
from a workplace standpoint time is money, saving token costs but wasting employee time with shittier model isn't a good enterprise use case. also, open source has its own risks and costs, anthropic is the best service offering overall factoring all that in
2026-08-17 16:16:58
0
биг боб :
nope
2026-08-17 10:07:28
0
To see more videos from user @lord.luke5, please go to the Tikwm
homepage.
