OpenAI's GPT-5.6 Sol autonomously post-trained the smaller Luna model with a "fairly underspecified prompt"

OpenAI has announced the launch of GPT-5.6 Sol, which independently fine-tuned the smaller Luna model in response to a fairly under-specified prompt. Sol achieved impressive results in OpenAI's internal RSI benchmark, scoring 16.2 points higher than GPT-5.5, showcasing the model's effective self-improvement capabilities. These developments mark a significant step towards OpenAI's goal of creating an automated researcher capable of operating independently, opening new horizons in the field of artificial intelligence.