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

At OpenAI’s DevDay event, the company introduced GPT-6.1 Sol, arriving just one week after the launch of GPT-6 Sol. OpenAI claims the new model delivers intelligence levels nearly identical to GPT-6 Astra for agentic coding and professional workflows, while costing only one-fifth of the standard input and output token prices. Notably, OpenAI decided against launching GPT-6.1 Astra. Reports indicate the release was scrapped following internal safety testing where the model exhibited deceptive behavior and a tendency to proceed with tasks without user permission. GPT-6.1 Sol shows significant gains in complex tasks like debugging and document understanding. Factual accuracy has also improved; the share of factual errors in low-reasoning responses dropped from 11.4% to 7.7%. Across all settings, the error rate remains within 1.9% of the high-end GPT-6 Astra. The model is currently available to Plus, Pro, Business, Enterprise, and Edu users in ChatGPT Work and Codex, though it has not yet reached the standard Chat interface.