docs: prompt caching antropic - #19253
Closed
Harshit28j wants to merge 1 commit into
Closed
Conversation
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
ghost
suggested changes
Jan 16, 2026
|
|
||
| ### 5. Verify Cache Hits | ||
|
|
||
| Check the LiteLLM logs to confirm routing: |
There was a problem hiding this comment.
@Harshit28j I meant the ui logs. The flow would be run query -> check ui to confirm it's sent to the correct model id
And have a screenshot showing the ui log
| litellm.router: Routing to deployment: deployment-abc-123 | ||
| ``` | ||
|
|
||
| Or check the LiteLLM UI: |
|
|
||
| # Second request: | ||
| litellm.router_utils.pre_call_checks.prompt_caching_deployment_check: Found cached model_id: deployment-abc-123 | ||
| litellm.router: Routing to deployment: deployment-abc-123 |
| For a prompt to be cached: | ||
| 1. **Minimum token count**: >1024 tokens (Anthropic requirement) | ||
| 2. **Cache control directive**: Must include `cache_control` in the message | ||
| 3. **Supported models**: Anthropic Claude models (3.5 Sonnet, 3 Opus, etc.) |
There was a problem hiding this comment.
We support more than just anthropic.
Is this ai generated? Seems odd to include this.
| - Routes to deployment A (`model_id: deployment-abc-123`) | ||
| - Cache hit! 90% cost reduction on input tokens | ||
|
|
||
| ## Performance Benefits |
There was a problem hiding this comment.
Please delete this. This is just ai slop.
|
|
||
| ### Q: Does this work with other providers besides Anthropic? | ||
|
|
||
| **A**: Currently, prompt caching routing is optimized for Anthropic Claude models. Other providers may be supported in the future. |
Contributor
|
Automated remediation patch bundle (generated due no direct push access to this PR branch).\nPlease cherry-pick/apply the patch proposal below and resolve corresponding threads.\n\n## PR #19253 — Unresolved threads
Proposed patchH1 — Replace ambiguous log checks with UI-log verification + screenshots diff --git a/docs/my-website/docs/tutorials/prompt_caching_routing.md b/docs/my-website/docs/tutorials/prompt_caching_routing.md
@@
-Check the LiteLLM logs to confirm routing:
+Validate routing from the LiteLLM UI Logs page:
+1. Run the same prompt twice.
+2. Open `/ui` → Logs.
+3. Confirm both requests resolved to the same `model_id`.
+4. Capture screenshots for first-call miss and second-call hit.
+
+
+H2 — Correct provider claim and FAQ wording @@
-3. **Supported models**: Anthropic Claude models (3.5 Sonnet, 3 Opus, etc.)
+3. **Supported models**: Any model/provider where prompt-caching metadata is supported in the request path.
@@
-**A**: Currently, prompt caching routing is optimized for Anthropic Claude models. Other providers may be supported in the future.
+**A**: Prompt caching routing is provider-agnostic; effective cache behavior depends on provider/model cache semantics.H3 — Remove low-signal performance marketing section @@
-## Performance Benefits
-### Cost Reduction
-... (remove speculative savings block) ... |
Contributor
|
This pull request has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This is a follow-up PR: #19233