GPT-6.1 Sol trails GPT-6 Sol on writing benchmark but nears Astra
On a writing benchmark, GPT-6.1 Sol's best setting scores 153 Elo below GPT-6 Sol and is neither cheaper nor faster, but it ties GPT-6 Astra at about a sixth of the price.
GPT-6.1 Sol seems like a step back for writing. On our writing benchmark, its best setting (xhigh) lands below GPT-6 Sol's best setting. That's 153 Elo lower, and it isn't cheaper or faster either on our tasks: $0.13 and 5.8 min per script, vs $0.11 and 2.9 min for GPT-6 Sol on max. OpenAI pitches 6.1 Sol as near-Astra performance at a lower cost, and for writing that holds. xhigh ties GPT-6 Astra on max (1976 vs 1968 Elo) for about a sixth of the price ($0.13 vs $0.85). So if you're on Astra for writing, switch. If you're on GPT-6 Sol, keep that one! More on every effort level 👇
This post quotes another post: x.com/openai/status/2104986129686741046