GPT-6.1 Sol

Released at DevDay on September 29, seven days after GPT-6 Sol, whose docs now point to it as the newer Sol model. OpenAI pitches it as "near-Astra performance for complex work at a lower cost": the same $2 / $10 per million input/output tokens as GPT-6 Sol, with cached input cut to $0.10, a 1.05M-token context, 128K output, and reasoning efforts from low to max (none and minimal are not supported). Artificial Analysis scores it 52 on Intelligence Index v4.3.2 at max effort, one point below GPT-6 Astra and four above GPT-6 Sol.

Like GPT-6 Astra, it is treated as Critical in cybersecurity under OpenAI's Preparedness Framework (High for biological and chemical capability), a step above GPT-6 Sol's High, so it ships on Astra's full safeguards stack. In a deployment simulation over 49,650 internal Codex tasks it drew 28 severity-3-or-higher flags (0.056%), against 27 for Astra, 42 for GPT-6 Sol and 63 for GPT-5.6 Sol. OpenAI says it matches Astra on DeepSWE v1.1 at roughly a fifth of the cost, and an Ultrafast tier of up to 300 tokens per second was announced alongside. At launch it is in ChatGPT Work and Codex for paid plans. The day before, OpenAI said it would not release the planned GPT-6.1 Astra because it fell short of the company's safety standards. Parameters undisclosed.