Independent tests put GPT-6.1 Sol's no-CoT results near Astra
A researcher's own runs show a large no-CoT jump for 6.1 Sol; if the looped-architecture hypothesis holds, monitorability is affected.
Researcher Rauno Arike ran GPT-6.1 Sol on 27 tasks and found it closes 80% of the no-CoT gap between GPT-6 Sol and GPT-6 Astra.
The runs reuse the task suite from an earlier Astra no-CoT evaluation, one sample per question; 6.1 Sol sits closer to Astra than to 6 Sol on 24 of 27 tasks. The 50% time-horizon estimate rises from 4 minutes for 6 Sol to 35 minutes, though most benchmarks saturate and the intervals are wide.
The author hypothesizes that 6.1 Sol is a looped transformer, possibly Astra Minor released under a different name, based on circumstantial evidence such as an Azure registry path; OpenAI has not confirmed this. If true, looped architectures are reaching below the frontier, with lower CoT controllability and monitorability as stated implications.
The results are one researcher's own runs, and the Astra and GPT-5.5 scores were taken from the earlier post rather than re-run.