OpenAI published a GPT-6.1 Sol safety addendum on September 29. Its introduction calls GPT-6.1 the latest model family in the GPT-6 series and describes Sol as offering capabilities comparable to GPT-6 Astra with a different speed-and-cost profile. OpenAI's developer model page lists GPT-6.1 Sol for API use and advises teams to compare it with Astra on their own tasks.
The addendum reports baseline safety evaluations. Under its Preparedness Framework, OpenAI says it treats GPT-6.1 Sol as Critical in cybersecurity and High in biological and chemical capability, and uses the safeguards stack described for GPT-6 Astra. These are the company's assessments, not independent performance or safety guarantees.
What to watch: developers can check the current model documentation for limits, pricing and account access, then test the model on their own workload before changing production systems. The cited materials establish this model introduction; they do not verify separate DevDay product launches or claims about other companies.