LiteLLM Responses API Drops cache_control on input_text Content Blocks

What this error means

LiteLLM Responses API silently drops cache_control on input_text content blocks during transformation is a LiteLLM / Anthropic API failure pattern reported for developers trying to fix litellm cache_control not forwarding to anthropic when using responses api. Based on the imported evidence, treat this as a tool-specific troubleshooting page rather than a generic API error.

Why this happens

Bug confirmed in litellm v1.83.14. In _transform_responses_api_content_to_chat_completion_content(), input_text handler builds output dict with only type and text, ignoring cache_control. Tools correctly preserve cache_control at lines ~1400-1401.

Common causes

• Prompt caching directives sent via Responses API are silently lost, causing higher API costs and slower responses. Tools correctly preserve cache_control but text content does not.
• Bug confirmed in litellm v1.83.14. In _transform_responses_api_content_to_chat_completion_content(), input_text handler builds output dict with only type and text, ignoring cache_control. Tools correctly preserve cache_control at lines ~1400-1401.

Quick fixes

1. Confirm the exact error signature matches LiteLLM Responses API silently drops cache_control on input_text content blocks during transformation.
2. Check the LiteLLM / Anthropic API 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/BerriAI/litellm/issues/27950

Evidence note: Bug confirmed in litellm v1.83.14. In _transform_responses_api_content_to_chat_completion_content(), input_text handler builds output dict with only type and text, ignoring cache_control. Tools correctly preserve cache_control at lines ~1400-1401.

Related errors

• LiteLLM cache_control on input_image dropped
• Anthropic API prompt cache miss

FAQ

What should I check first?

Start with the exact LiteLLM Responses API silently drops cache_control on input_text content blocks during transformation text and the smallest action that reproduces it.

Can I ignore this error?

No. Treat it as a failed LiteLLM / Anthropic API 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 LiteLLM Responses API silently drops cache_control on input_text content blocks during transformation.