OpenAI Launches GPT-6.1 Sol, Matching GPT-6 Astra Performance at Lower Cost

OpenAI has launched GPT-6.1 Sol, which is reported to deliver performance nearly matching GPT-6 Astra while costing one-fifth of the standard input and output token prices. The new model, showcased during OpenAI's DevDay event, achieves a factual error rate of 7.7% at low reasoning effort, a significant improvement from the 11.4% error rate of its predecessor, GPT-6 Sol. Notably, GPT-6.1 Sol's error rate remains within 1.9% of GPT-6 Astra, indicating its competitive performance. OpenAI claims that GPT-6.1 Sol demonstrates significant improvements over GPT-6 Sol, failing less often in challenging evaluations. This follows the recent launch of GPT-6 Sol just one week prior to the introduction of GPT-6.1 Sol. OpenAI also reported observing no attempts to bypass the automated safety reviewer with the new model.