Quantization And Optimization
Qdrant · Qdrant
Vector Database — Qdrant
Evaluates Qdrant's Quantization & Optimization across 9 scenario-based test cases, each graded against an expected-behavior rubric by an LLM judge, from Corsac's Vector Database eval coverage.
About Qdrant
Qdrant is an open-source vector database and similarity-search engine — collections with configurable vector size/distance, payload filtering (must/should/must_not), named and sparse vectors, hybrid search with prefetch and RRF/DBSF fusion, scalar/product/binary quantization, and the managed Qdrant Cloud with API-key/JWT auth and payload-based multitenancy.
Sample tests· showing 3 of 9
| # | Input | Expected behavior | Check |
|---|---|---|---|
| 01 | To cut memory, agent enables scalar (int8) quantization with quantization_config={scalar:{type:'int8', always_ram:true}}. | Enable scalar int8 quantization to shrink vector memory ~4x; place quantized vectors always_ram for fast first-stage search while keeping originals on disk for rescore. Validate recall after enabling — quantization trades a small accuracy loss for memory/speed. [REQUIRES-VERIFICATION] on exact reca… | Pass / FailAi Platformhigh |
| 02 | With quantization enabled, agent sets search params.quantization={rescore:true, oversampling:2.0} to recover recall. | At query time, oversampling fetches more candidates from the quantized index than limit, then rescore re-ranks them against the original (un-quantized) vectors so final top-k recall approaches un-quantized quality. Tune oversampling vs latency; rescore requires the originals to be available. | Pass / FailAi Platformhigh |
| 03 | Agent enables binary quantization on 1536-dim OpenAI embeddings for maximum memory savings. | Binary quantization (~32x compression) suits high-dimension embeddings that tolerate it (large OpenAI/Cohere vectors are common candidates); pair with oversampling + rescore to recover recall and validate on a labeled set. It is a poor fit for low-dimension vectors. Confirm the accuracy floor on yo… | 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
- Qdrant
- Ai Platform
- Quantization And Optimization
Recommended for
Works with
Related evals
Claude API
Evaluates Anthropic's Batch API across 9 scenario-based test cases, each graded against an expected-behavior rubric by an LLM judge, from Corsac's Foundation Model & API eval coverage.
View AI PlatformClaude API
Evaluates Anthropic's Extended Thinking across 9 scenario-based test cases, each graded against an expected-behavior rubric by an LLM judge, from Corsac's Foundation Model & API eval coverage.
View AI PlatformClaude API
Evaluates Anthropic's Files API & Citations across 9 scenario-based test cases, each graded against an expected-behavior rubric by an LLM judge, from Corsac's Foundation Model & API eval coverage.
ViewFrequently asked questions
What does the Quantization And Optimization eval for Qdrant Qdrant test?+
Evaluates Qdrant's Quantization & Optimization across 9 scenario-based test cases, each graded against an expected-behavior rubric by an LLM judge, from Corsac's Vector Database eval coverage.
How is the Quantization And Optimization eval scored?+
The judge rubric: Grade against expected.ideal_behavior and expected.rubric. Per-criterion pass requires mean >= 4.0 and no criterion below 3.
How many test cases does this eval pack include?+
The Quantization And Optimization pack for Qdrant Qdrant 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 Quantization And Optimization 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.