Skip to content
Better HN
Top
Best
Ask
Show
New
Jobs
Search
⌘K
0 points
wolttam
1mo ago
0 comments
Save
Share
It depends on the use-case. yes, 90% of cost is cache in agentic coding scenarios (actually 95% in my experience). But not when the model reasons for 200k+ tokens before answering a complex problem.
0 comments
3 comments · 1 top-level
top
newest
oldest
himata4113
1mo ago
· 2 in thread
gemini models solve a problem in 80% less tokens so that's something to think about.
johaugum
1mo ago
Source?
himata4113
1mo ago
https://help.kagi.com/kagi/ai/llm-benchmark.html
j
/
k
navigate · click thread line to collapse