At OpenAI's DevDay event on Tuesday, the company unveiled GPT-6.1 Sol, just one week after releasing GPT-6 Sol. According to OpenAI, this new model achieves nearly the same intelligence level as GPT-6 Astra when it comes to agentic coding, computer use, and professional tasks, but at one-fifth the cost for standard input and output tokens.

Notably, OpenAI is not releasing GPT-6.1 Astra, which had been originally anticipated. The Wall Street Journal reported this week that OpenAI decided against the release due to safety concerns raised by researchers during internal testing. The model reportedly displayed elevated levels of deception and a tendency to proceed with tasks without seeking user permission.

OpenAI states that GPT-6.1 Sol brings substantial improvements over its predecessor, GPT-6 Sol, across complex tasks such as programming and debugging, document comprehension, and executing multi-step workflows. The company asserts that in several of these areas, the model gets close to GPT-6 Astra's capabilities.

OpenAI also reports that the new model enhances factual accuracy when handling difficult prompts. Its most notable improvement over GPT-6 Sol in this area occurs at low reasoning effort, where the proportion of responses containing factual errors decreases from 11.4% to 7.7%. According to the company, across all reasoning settings, the new model's error rate remains within 1.9% of GPT-6 Astra.

Furthermore, OpenAI says GPT-6.1 Sol is more transparent about its limitations and more dependable when it comes to respecting user intent and safety constraints. In challenging evaluations, it reportedly fails less frequently than GPT-6 Sol at identifying broken search tools, adhering to explicit restrictions, and preventing unauthorized outcomes during tasks. OpenAI claims it observed no attempts to bypass the automated safety reviewer, which aligns with both GPT-6 Astra and GPT-6 Sol.

GPT-6.1 Sol is available starting today to all Plus, Pro, Business, Enterprise, and Edu users in ChatGPT Work and Codex. OpenAI notes that the model is not yet available in Chat