Files Api And Citations
Claude API · Anthropic
Foundation Model & API — Anthropic (Claude)
Anthropic evals — Files API & Citations (relift v3 InfraRed)
About Anthropic
Anthropic is an AI safety company and the maker of Claude. Its API exposes the Claude model family (Opus, Sonnet, Haiku) with tool use, prompt caching, extended thinking, batch processing, vision, the Files and Memory tools, and the Claude Agent SDK.
Sample tests· showing 3 of 9
| # | Input | Expected behavior | Check |
|---|---|---|---|
| 01 | Operator uploads a 12 MB invoice PDF via POST /v1/files and receives a file_id. | Persist file_id with its source-document mapping. Reference it in Messages calls via source={type:'file', file_id:'<id>'} on a document content block. Do not re-upload identical files (use prior file_id) — but plan for deletion lifecycle. | Pass / FailAi Platformmedium |
| 02 | Operator wants the model to answer with citations from a previously-uploaded research PDF (file_id=file_abc). | Include a document content block with source={type:'file', file_id:'file_abc'} and citations={enabled:true}. The response will include citation blocks tied to specific char locations or page numbers within the document. | Pass / FailAi Platformhigh |
| 03 | Response contains citation blocks: cited_text, document_index, document_title, char_location with start_char_index / end_char_index. | Render each citation as a deep link to the source. Preserve cited_text verbatim. Map document_index back to the order of documents in the request to attribute correctly. Do not paraphrase cited_text in the rendered citation. | Pass / FailAi Platformhigh |
How this eval is graded
Grade against expected.ideal_behavior and expected.rubric. Per-criterion pass requires mean >= 4.0 and no criterion below 3.
Rubric criteria
- Anthropic
- Ai Platform
- Files Api And Citations
Recommended for
Works with
Related evals
Run this eval in your workspace
Connect your data, configure thresholds, and review results with your team.