codex - 💡(How to fix) Fix Codex consumed 19G RAM [1 comments, 2 participants]

TL;DR

The issue of Codex consuming excessive RAM (19G) for an extended period (30 minutes) may be related to the specific model (gpt-5.5) or the input data, and investigating the input or model configuration could help mitigate the issue.

Guidance

• Review the uploaded thread (019dfae5-1cac-70d2-9883-2e355436edb9) to understand the input data and model usage that led to high RAM consumption.
• Check if there are any known issues or optimizations for the gpt-5.5 model that could be contributing to the high memory usage.
• Consider testing with a different model or input data to isolate if the issue is model-specific or data-dependent.
• Monitor system resources during the execution to identify any patterns or spikes in memory usage that could indicate the root cause.

Notes

The provided information does not include details about the specific commands or parameters used with the Codex CLI, which could be crucial in diagnosing the issue. Additionally, the version of Codex CLI (0.128.0) and the PRO subscription might have implications on the available features or support, but without further context, it's difficult to assess their relevance to the problem.

Recommendation

Apply workaround: Given the lack of detailed information about the commands or parameters used, and without a clear indication of a version-specific issue, applying a workaround such as optimizing input data, testing with different models, or adjusting usage patterns might be the most feasible next step.