WHAT NEW IN X10.31 SAT

AI collections from X · 22 collections · 61,986 posts

Back to No.003 GPT-6 Sol leaks, screenshots and release rumorsUpdated 12 days ago
gpt-6IMAGE

NYT Connections benchmark: Sol slightly weaker but much cheaper

Lech Mazur's Extended NYT Connections Benchmark shows Opus 5.5, GPT-6 Sol and Luna all slightly behind their predecessors but far cheaper, with data on GitHub.

Lech Mazur
@LechMazurArchived 2026.09.24Originally posted 2026.09.23
NYT Connections benchmark: Sol slightly weaker but much cheaper (image 1 of 2)
NYT Connections benchmark: Sol slightly weaker but much cheaper (image 2 of 2)
Lech Mazur @LechMazurOriginally posted 2026.09.23
Open original on X

Opus 5.5, GPT-6 Sol, and GPT-6 Luna are all slightly worse but a lot cheaper than their predecessors on the Extended NYT Connections Benchmark. More info: https://github.com/lechmazur/nyt-connections/

x.com/lechmazur/status/2102536776799629717Archived 2026.09.24 10:07
Open original on X