Search Mode Selection Query Routing
Exa · Exa Labs
Exa Labs evals — Search Mode Selection & Query Routing (relift v3)
About Exa Labs
Exa (formerly Metaphor) is a neural search API that understands the meaning of queries rather than matching keywords — returning the most relevant URLs and content from the web for any semantic question. Developers use Exa to power research agents, content discovery, and RAG pipelines.
Sample tests· showing 3 of 10
| # | Input | Expected behavior | Check |
|---|---|---|---|
| 01 | Voice assistant needs lexical SKU lookup from public retailer pages; latency budget under 200ms; no synthesis. | Choose type=instant (or fast) for latency-sensitive lexical lookups; reserve deep for multi-hop research with additionalQueries and synthesis. | Pass / FailRoutinghigh |
| 02 | Research agent queries 'transformer efficiency arxiv 2026' via POST /search. | Use category=research paper (not news) and date filters; explain category hint steers corpus, not query keywords alone. | Pass / FailRoutingmedium |
| 03 | Mobile app needs <500ms median with slightly richer ranking than instant; no outputSchema. | Contrast fast vs instant latency/quality tradeoff per docs; pick fast when instant quality is insufficient but deep is too slow. | Pass / FailRoutinglow |
Rubric criteria
- Exa Labs
- Search
- Search Mode Selection Query Routing
Recommended for
Works with
Related evals
Run this eval in your workspace
Connect your data, configure thresholds, and review results with your team.