Codex produces weak explanations compared to task execution quality and other models

What version of Codex is running?

codex-cli 0.72.0

What subscription do you have?

ChatGPT Pro

Which model were you using?

gpt-5.2 xhigh

What platform is your computer?

Darwin 25.2.0 arm64 arm

What issue are you seeing?

Summary

Codex performs task execution very well, especially for coding and refactoring, but its explanations are consistently weaker than expected.

Details

When Codex completes a task, the explanation of what it did and why is often shallow, incomplete, or poorly articulated. This makes it harder to understand the changes, even though the final output itself is usually correct.

More generally, when I ask questions about an existing codebase (for example, “explain this module,” “why is this implemented this way,” or “walk me through this logic”), Codex’s explanations are noticeably less clear, less structured, and less insightful compared to Claude.

Expected behavior

• Clear, step-by-step explanations of changes made
• Better articulation of reasoning (intent, trade-offs, assumptions)
• Stronger explanations when answering questions about a codebase, not just task completion

Actual behavior

• Tasks are executed correctly
• Explanations lack depth, structure, and clarity
• Reasoning is often implicit or missing

Impact

This reduces Codex’s usefulness as a learning and reasoning companion, especially when working with unfamiliar or complex codebases.

What steps can reproduce the bug?

Uploaded thread: 019b2b6f-e691-7ef1-8c56-47c445e72ca0

What is the expected behavior?

Expected behavior

• Clear, step-by-step explanations of changes made
• Better articulation of reasoning (intent, trade-offs, assumptions)
• Stronger explanations when answering questions about a codebase, not just task completion