turbopuffer
For turbopufferData AnalysisVector Db

Vector Schema And Distance

turbopuffer · turbopuffer

Serverless vector database — turbopuffer

Evaluates turbopuffer's Vector Schema & Distance across 9 scenario-based test cases, each graded against an expected-behavior rubric by an LLM judge, from Corsac's Serverless vector database eval coverage.

About turbopuffer

turbopuffer is a serverless vector database designed for high-performance approximate nearest neighbor search at scale. It handles ingest, indexing, and hybrid queries with a simple HTTP API and charges only for storage and queries — no always-on infrastructure.

Employees

~10

Industry

Vector Database

Headquarters

United States

Sample tests· showing 3 of 9

#InputExpected behaviorCheck
01

Model upgrade shipped wrong embedding size to prod namespace legal-emb-768.

Reject or surface dimension mismatch on write; do not silently truncate; fix embedding pipeline before upsert.

Pass / FailSchemacritical
02

768-dim vectors indexed; new model outputs 1024-dim; team asks to hot-swap query vectors only.

Plan new namespace or full re-upsert with consistent dimension; run recall@k per /docs/recall; do not invent hot-swap metric API.

Pass / FailSchemahigh
03

Latency SLO 40ms p99; recall target 0.95 [REQUIRES-VERIFICATION] for tuning knobs.

Contrast ANN approximate vs exact kNN per query docs; mark tuning numbers [REQUIRES-VERIFICATION]; pick ANN for scale with recall validation.

Pass / FailPerformancemedium

Unlock full benchmark

6 more test cases

Use this benchmark

How this eval is graded

Grade against expected.ideal_behavior and expected.rubric. Penalize failure_modes.

Rubric criteria

  • Turbopuffer
  • Vector Db
  • Vector Schema And Distance

Recommended for

turbopufferturbopuffer customers

Works with

Related evals

Frequently asked questions

What does the Vector Schema And Distance eval for turbopuffer turbopuffer test?+

Evaluates turbopuffer's Vector Schema & Distance across 9 scenario-based test cases, each graded against an expected-behavior rubric by an LLM judge, from Corsac's Serverless vector database eval coverage.

How is the Vector Schema And Distance eval scored?+

The judge rubric: Grade against expected.ideal_behavior and expected.rubric. Penalize failure_modes.

How many test cases does this eval pack include?+

The Vector Schema And Distance pack for turbopuffer turbopuffer contains 9 test cases. 3 sample cases are shown free on this page; the full set runs in a Corsac workspace.

How do I run this eval?+

Sign up for Corsac, connect your model or agent endpoint, and run the Vector Schema And Distance pack as-is or after customizing thresholds. Results land in your workspace with per-case scores, and you can gate releases on the pack in CI via the REST API.

Run this eval in your workspace

Connect your data, configure thresholds, and review results with your team.