OpenAI has introduced GPT-6.1 Sol, a significant upgrade to its previous GPT-6 Sol model, which brings capabilities closer to those of the flagship GPT-6 Astra. Remarkably, GPT-6.1 Sol is priced five times lower than Astra, with costs set at $2 for every million input tokens and $10 for million output tokens. In benchmarks like DeepSWE v1.1, which assesses the execution of complex programming tasks within real codebases, GPT-6.1 Sol's performance is on par with GPT-6 Astra. Moreover, in the GDP.pdf test, which measures the accuracy of model responses to complex professional inquiries using PDF documents, GPT-6.1 Sol outperforms Opus 5.5. OpenAI has also reported a decrease in the share of factual errors in GPT-6 Sol's responses, dropping from 11.4% to 7.7%. Starting today, GPT-6.1 Sol will be available to users subscribed to Plus, Pro, Business, Enterprise, and Edu plans in ChatGPT Work and Codex, with plans to roll it out to the ChatGPT chatbot later. This release positions OpenAI strongly in the market, potentially reshaping competition by offering advanced features at a more accessible price point.
Informational material. 18+.
