gpt-6IMAGE

Four models write MathArena blog posts, only Opus-5.5 avoids slop

Given 48 hours and a $500 API budget to write a math blog post, GPT-6 Astra, GPT-6.1 Sol and GLM-5.3 produced slop, while Opus-5.5 wrote a good post with interesting conclusions.

Jasper Dekoninck

@j_dekoninckArchived 24 days agoOriginally posted 2026.10.08

Four models write MathArena blog posts, only Opus-5.5 avoids slop

Jasper Dekoninck @j_dekoninckOriginally posted 2026.10.08

Open original on X

Can LLMs write interesting and novel blog posts for MathArena? We give four agents 48 hours and a $500 API key to create an interesting MathArena blog post. GPT-6 Astra, GPT-6.1 Sol, and GLM-5.3 produced slop. Opus-5.5 created a good blog post with interesting conclusions.

x.com/j_dekoninck/status/2108218750197796867Archived 2026.10.09 10:38