support prompt caching for claude models to cut latency and api costs#2348
Open
thechaitanyaanand wants to merge 2 commits into
Open
support prompt caching for claude models to cut latency and api costs#2348thechaitanyaanand wants to merge 2 commits into
thechaitanyaanand wants to merge 2 commits into
Conversation
thechaitanyaanand
changed the title
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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 PR adds support for ephemeral prompt caching for anthropic models (claude) within the model interaction layer.
The Problem: When running multi-step agent loops (like ReAct), the system prompt, available tools, and message history get sent back and forth to the model on every single step. This leads to redundant token processing, which slows down response times and increases costs.
Where it matters: For longer runs or complex agents with many tools, this can cut down token latency and costs by up to 90% since Claude can retrieve the prompt context from its cache instantly instead of reading it again.
Added a clean unit test in
tests/test_models.py(test_prepare_completion_kwargs_prompt_caching) to ensure the headers are formatted correctly for Claude, and completely ignored for other models.