OpenAI GPT-6.1 Sol Cuts Astra Cost by 81% to Lead Efficiency Frontier

← Back to live feed · 1 stories across 1 day The latest AI release from OpenAI delivers intelligence within 1.04 points of GPT-6 Astra in the Agent Arena while reducing the median cost per task to $0.56. The GPT-6.1 Sol (Max) model currently ranks fifth in that arena, outperforming GPT-6 Sol by 1.52 points at a 39% lower cost. In the Code Arena for WebDev, the model initially took third place before sliding to fourth after Anthropic released Claude Sonnet 5.5. OpenAI introduced the model during its DevDay event, pricing it at $2 per million input tokens and $10 per million output tokens. This version replaces GPT-6 Sol just seven days after that launch and outperforms Anthropic's Fable 5.1 and Opus 5.5 in agent tasks at 88% and 65% lower costs, respectively. While Sol targets the cost efficiency Pareto frontier, Google's Gemini 4 Argon (High) now holds the top spot in the Text Arena.