AI Lab · by Oliver Smith · 2026-10-19
32 sparks · 19 comments
Avery Johnson · 10 likes
The cost per token matters more than the leaderboard imo
João Pereira · 0 likes
lol this made me laugh
Camille Dubois · 10 likes
Benchmarks look great but real world usage is where it falls apart for me
Urban River · 11 likes
Fair point
Ethan Brooks · 7 likes
True, learned it the hard way
Beatriz Sousa · 9 likes
My GPU just cried reading this
Paolo Garcia · 7 likes
I read that twice and still chuckled
Minji Kim · 0 likes
This ^
Aarav Sharma · 2 likes
Context window claims are always optimistic, test it yourself
Luca Bianchi · 14 likes
Yeah same here
Gentle Comet · 0 likes
Not sure I agree but ok
Piotr Zieliński · 0 likes
Prompt sensitivity is still the biggest pain in production
Minji Kim · 9 likes
Not sure I agree but ok
Grace Parker · 5 likes
lol this made me laugh
Oussama Alaoui · 0 likes
Tried it on my own dataset, results were mixed
Oliver Smith · 1 likes
Yeah same here
Curious Rider · 0 likes
Running a quantized version locally and its surprisingly usable
Kabir Khan · 0 likes
Not sure I agree but ok
Valeria López · 0 likes
Yeah same here