FILTERALL 3■ BUILD 1◇ RESEARCH 2▤ FRAMEWORK 0
all projects
◇ RESEARCH● PUBLISHED
Reasoning benchmark — What does reasoning actually cost in your token budget? A benchmark of 13 frontier models across reasoning economy, legibility, and correctness
Reasoning is 80% of a token budget. This project documents how it is spent.
#reasoning#benchmark#tokenomics
4 ARTIFACTS
View project →
◇ RESEARCH● PUBLISHED
Token Language Tax — How non-english language is taxed by tokenisers across models
Danish costs 50% more than English to say the same thing. This project measures the surcharge across 7 models and 11 languages
#Tokenomics#Sovereign AI#Benchmark
2 ARTIFACTS · 1 ESSAY
View project →
/ project philosophy
“Every Nowable project has a learning objective. I use technology as means to demystify AI because it helps me understand the applications and architectures we use every day. ”