GPT-5.6 Sol Can Now Handle Much Bigger Coding Projects in Codex

OpenAI has expanded the capabilities of GPT-5.6 Sol in Codex, enabling ChatGPT subscribers to process larger amounts of code and conversation history within a single coding session. The model now supports a 1.05 million-token context window, which essentially refers to the amount of information the AI can retain simultaneously, including prompts, source code, files, previous messages, and tool results.

Previously, this full context length was only accessible to Codex users through API-key usage, but OpenAI has now enabled it for ChatGPT accounts as well. This additional room can prove beneficial when handling extensive software projects, as Codex can retain more files, prior changes, command outputs, and conversation history before it needs to compress older information.

This could be especially helpful for large codebase refactoring, complex debugging, and long-running development sessions where losing earlier context could diminish the AI coding assistant's effectiveness.

GPT-5.6 Sol is available to Codex users on supported ChatGPT plans, including Plus and Pro. However, it's important to note that the larger context window does not equate to a larger usage allowance. OpenAI estimates that GPT-5.6 Sol allows for around 10 to 100 local Codex messages per five-hour period on Plus plans, although the actual usage can vary significantly.

Additional weekly limits may also apply. Therefore, while the larger context window provides more room when a difficult project genuinely requires it, enabling such a large context for routine coding could result in available usage disappearing considerably faster.