Codex Context Compaction Fails with Unknown Parameter 'prompt_cache_retention'

What this error means

Unknown parameter: 'prompt_cache_retention' is a Codex failure pattern reported for developers trying to fix codex context compaction error unknown parameter prompt_cache_retention. Based on the imported evidence, treat this as a tool-specific troubleshooting page rather than a generic API error.

Why this happens

38 comments on #17809. Error occurs in both macOS app and CLI. API returns invalid_request_error with code ‘unknown_parameter’ for prompt_cache_retention. Also affects third-party proxies (new-api #4462). Affects ChatGPT Pro users.

Common causes

• Codex is OpenAI’s paid coding tool (ChatGPT Pro/Plus). Context compaction is a core feature that breaks on all threads, preventing long sessions. The error is an API-level incompatibility that users cannot fix themselves.
• 38 comments on #17809. Error occurs in both macOS app and CLI. API returns invalid_request_error with code ‘unknown_parameter’ for prompt_cache_retention. Also affects third-party proxies (new-api #4462). Affects ChatGPT Pro users.

Quick fixes

1. Confirm the exact error signature matches Unknown parameter: 'prompt_cache_retention'.
2. Check the Codex account, local tool state, and provider configuration involved in the failing workflow.
3. Compare the failing environment with a known working setup, then change one configuration value at a time.

Platform/tool-specific checks

• Verify the command, editor, extension, or API client that produced the error.
• Compare local settings with CI, deployment, or editor-level settings when the error appears in only one environment.
• Avoid deleting credentials, local model data, or project settings until the failing scope is clear.

Step-by-step troubleshooting

1. Capture the exact error message and the command, editor action, or request that triggered it.
2. Check whether the failure is account/auth, quota/rate, model/provider, local runtime, or deployment configuration.
3. Review the source evidence below and compare it with your environment.
4. Apply one change at a time and rerun the smallest failing action.
5. Keep the working fix documented for the team or deployment environment.

How to prevent it

• Keep provider/tool configuration documented.
• Record non-secret diagnostics such as tool version, provider name, model name, and command path.
• Add a lightweight check before CI or production workflows depend on the tool.

Sources checked

• https://github.com/openai/codex/issues/17809
• https://github.com/openai/codex/issues/20931
• https://github.com/QuantumNous/new-api/issues/4462

Evidence note: 38 comments on #17809. Error occurs in both macOS app and CLI. API returns invalid_request_error with code ‘unknown_parameter’ for prompt_cache_retention. Also affects third-party proxies (new-api #4462). Affects ChatGPT Pro users.

Related errors

• Codex Error running remote compact task
• Codex stream disconnected before completion
• Codex Unknown parameter client_metadata

FAQ

What should I check first?

Start with the exact Unknown parameter: 'prompt_cache_retention' text and the smallest action that reproduces it.

Can I ignore this error?

No. Treat it as a failed Codex workflow until the root cause is understood.

Is this guaranteed to have one fix?

No. The imported evidence supports the troubleshooting path above, but tool behavior can vary by account, plan, version, provider, and local configuration.

How do I know the fix worked?

Rerun the same command, editor action, or request. The fix is working when that action completes without Unknown parameter: 'prompt_cache_retention'.