OpenAI launches GPT-6.1 Sol, approaches Astra capabilities
OpenAI has introduced its new AI model, GPT-6.1 Sol, which enhances coding abilities and factual accuracy. The model is now available to select ChatGPT users.
AI company OpenAI has launched its latest language model, GPT-6.1 Sol, highlighting its enhanced capabilities.
At its DevDay event, the company announced that GPT-6.1 Sol approaches GPT-6 Astra in agentic capabilities, computer use, and professional tasks. However, it comes at one-fifth the cost of Astra's standard token prices. The new model also improves programming, debugging, document understanding, and multi-step workflows.
GPT-6.1 Sol is available from September 29 to Plus, Pro, Business, Enterprise, and Edu users through ChatGPT Work and Codex. The model has not yet been integrated into Chat.
The release follows a week after the introduction of GPT-6 Sol. Reports indicated that the previously planned GPT-6.1 Astra was scrapped due to safety concerns, particularly regarding its tendency to act without user permission.
According to OpenAI, GPT-6.1 Sol also improves factual accuracy on difficult prompts. At low reasoning effort, the share of responses with factual errors fell from 11.4% to 7.7%.