Sonnet 5.5 is a bit worse at this test - physics seem a bit less real & more LoCs. But the shocker is token use at xHigh and Max: xhigh: Opus 52.5k vs Sonnet 427.8k Max: Opus 175.3k vs Sonnet 460.9k This is just one test, so could be a fluke, but kind of crazy
Benchmark idea: Who can use fewest lines of code to do the same thing? In this test, Opus 5.5 uses ~half the lines of code that of Astra and the physics/visual quality isn't noticeably worse. I am trying to approximate how 'elegant' the code is - something that many complain in AI-generated code. While isn't perfectly true, fewer lines of code could mean more elegant, generalised approach.

Sep 29, 2026 · 7:40 AM UTC

4
1
59
6,572
Sort replies: Relevant Recent Liked
Replying to @petergostev
How opus made it so cheap?
1
34
Opus is still more expensive than Astra
1
81
Replying to @petergostev
Normal. Give a 100 and 130iq dev a complex task, iq130 wins, always, even at double cost. Simple task: 100iq wins because cheaper.
21
Replying to @petergostev
Interesting perspective
29
Replying to @petergostev
Did you compare dollar cost? Sonnet is cheaper per token, curious whether 427k tokens at xhigh still ends up costing more than Opus's 52k.
1
210