Why does this matter for AI cost planning?

Managed vs self-hosted can differ 10-20x for the same workload Pricing models vary wildly — per-vector, per-pod, per-dimension, per-cluster Preset workload sizes (100K, 1M, 50M, 1B vectors) show you where cost scales break

Vector DB Cost · for RAG builders & CTOs

Pinecone vs pgvector vs Qdrant vs 5 more

Managed vs self-hosted. Serverless vs pod-based. Price differences up to 20x for the same workload.

Pricing verified: 2026-09-11 8 vendors compared

What this calculator does

Compare 8 vector databases (Pinecone, Weaviate, Qdrant, pgvector, Chroma, and more) at your exact workload.

Why use it

Managed vs self-hosted can differ 10-20x for the same workload
Pricing models vary wildly — per-vector, per-pod, per-dimension, per-cluster
Preset workload sizes (100K, 1M, 50M, 1B vectors) show you where cost scales break

New to this calculator? Start with the ⚡ Playground — a few sliders, instant ballpark. Then switch to the 🧮 Calculator for your exact number.

Two ways to use this: visualize in the Playground, then get your number in the Calculator.

▶Playground A quick, visual way to see which factors move your result the most. Open the playground → ▤Calculator Enter your real workload for a precise result you can apply to your own usage. Go to the calculator →

What will a vector database cost?

Storage scales with vectors × dimensions; queries add ongoing cost. Provider fixed at default.

Estimated monthly cost

—

Change the input sliders below to see new estimates.

Vectors 5,000,000

Queries / mo 1,000,000

How we got this estimate

—

💡Cost = vector storage (count × dimensions) + query volume. Higher dimensions mean more bytes per vector.

Pinecone, pgvector, Qdrant and more, priced side by side. Enter your corpus below for an honest comparison.

📊 Not sure of a value? Fields with a ▾ Typical pill offer broad industry ballparks (sourced typical ranges) so you can move forward now — your result gets more accurate as you replace them with your own measured numbers. Values marked * are rough estimates.

🎮 Interactive Guide & Calculator Playground →

Enter key values

Total vectors (chunks × docs), dimensions, QPS peak, writes/month.

Or load a preset

Early-stage (100K), Mid-stage (1M), Enterprise (50M), Billion-scale — each loads typical numbers.

Read the winner

Cheapest option at your workload is highlighted. Managed vs self-host delta shown separately.

Act on it

At small scale, managed wins on operational cost. At billion-scale, self-host becomes unbeatable.

Enter workload

Total vectors = (corpus docs × chunks per doc). E.g., 10K docs × 3 chunks = 30K vectors. Dimensions — 1024 typical for Voyage/Cohere, 1536 for OpenAI 3-small, 3072 for OpenAI 3-large. Higher dimensions cost more on some DBs (Weaviate is per-dimension). QPS peak = queries per second during peak traffic. Writes/month = new vectors added each month.

Pick deployment preference

Managed (Pinecone, Weaviate Cloud, Qdrant Cloud) = no ops burden but higher cost. Self-hosted (pgvector, Qdrant self-host, Milvus) = lower cost but you operate it. "Any" shows all 8 side-by-side. At early scale, the operational time to self-host eats more than the managed premium. At billion-scale, self-host becomes unavoidable on cost.

Compare all 8

Results table shows monthly + annual cost across all 8 vendors sorted cheapest first. Gold highlight = your current selection. Green = cheapest. Managed vs self-hosted delta is shown — useful for the "is the ops burden worth it" decision. Each vendor's pricing model (per-vector, per-pod, per-dimension, etc.) is shown in the pricing model column.

Next actions

This covers the DB layer only. For full RAG cost: RAG Pipeline Cost. For embedding model choice: Embedding Cost. For self-host break-even at LLM layer: Self-Host Break-even. If your corpus is growing, re-run this quarterly — you often cross scale boundaries and should switch DBs.

👇 Now try the calculator below with your own AI workloads

🎛 CALCULATOR

📊 Your vector workload

We'll compute monthly cost across all 8 options.

Total vectors Hint: How many chunks/embeddings you'll store. 10K docs x 3 chunks ~ 30K.

Vector dimensions Hint: Size of each search fingerprint. Higher = more storage/RAM. Leave default unless you know it.

Queries per second (peak) Hint: Searches per second at peak. 5/sec ~ 13M a month.

Vector writes per month Hint: New vectors added each month.

Deployment preference Hint: Managed service vs self-hosted. Drives the pricing model.

Load preset

📈 RESULTS

Cheapest option at your workload

Cheapest managed

Cheapest self-hosted

Managed vs self-host

💡 Recommendations

📋 All 8 vector DB options ranked

Same workload, different vendors. Green = cheapest, gold highlight = your current selection.

Vector DB	Pricing model	Monthly cost	Annual

Embedding model cost → RAG vs Fine-Tuning → Get a vector DB architecture review →

🎯 Use this result to

🗂️ Pick the right vector DB — Pinecone wins under 1M vectors, others at scale. Get a concrete answer for your numbers.
📊 See full TCO not sticker — Storage + queries + write costs + managed-service premiums all included.
⛰️ Spot scale cliffs — Pricing tiers flip at 5M, 10M, 50M vectors. Calc surfaces them before you commit.
🔌 Integrate with your AI agents — MCP available for agentic workflow integration. Cost-aware vector retrieval.

📅 Schedule a call to apply this to your workload

📋 What now?

Managed vs self-hosted is the big lever — the delta card shows it; below ~1–2M vectors managed usually wins on total cost once you price your ops time.
Dimensions drive storage — if recall holds, a 1024-dim model over 3072 can cut index size (and some per-dimension DB bills) by roughly two-thirds.
Size on peak QPS, not average — managed tiers step up on peak throughput; confirm your real peak before committing to a tier.

📅 Book a working session to apply this to your workload →

Need help using this calculator for your workloads?

AICost.ai has 96+ calculators and playbooks. Schedule an AvatarVA meeting and we'll work through your real cost scenarios across AI & Cloud: visibility, cost reduction, optimization, forecasting and capacity planning, without sacrificing accuracy or performance.

📅 Schedule an AvatarVA meeting →

Vendor / Model	Field	Why it’s inferred
Anthropic — Claude Sonnet 4.6	`cachedInput`	Derived at 10% of input rate — Anthropic publishes 90% cache-hit discount on this tier.
Anthropic — Claude Sonnet 4.5	`cachedInput`	Derived at 10% of input rate; same 90% cache-hit convention as Sonnet 4.6.
Anthropic — Claude Sonnet 4.5	`batchInput`	Derived at 50% of standard input — Anthropic documents uniform 50% Batch discount.
Anthropic — Claude Sonnet 4.5	`batchOutput`	Derived at 50% of standard output — Anthropic documents uniform 50% Batch discount.
Anthropic — Claude Haiku 4.5	`cachedInput`	Derived at 10% of input rate — Anthropic 90% cache-hit discount convention.
OpenAI — GPT-5.4 Mini	`cachedInput`	Derived at 10% of input — OpenAI documents automatic 90% discount on cache hits across GPT-5.x tier.
OpenAI — GPT-5.4 Nano	`cachedInput`	Derived at 10% of input — OpenAI 90% cache-hit convention.
OpenAI — GPT-5.4 Nano	`batchInput`	Derived at 50% of input — OpenAI Batch API uniform 50% discount.
OpenAI — GPT-5.4 Nano	`batchOutput`	Derived at 50% of output — OpenAI Batch API uniform 50% discount.
OpenAI — GPT-5.4 Pro	`cachedInput`	Derived at 10% of input — OpenAI 90% cache-hit convention.
OpenAI — GPT-5.4 Pro	`batchInput`	Derived at 50% of input — OpenAI Batch API uniform 50% discount.
OpenAI — GPT-5.4 Pro	`batchOutput`	Derived at 50% of output — OpenAI Batch API uniform 50% discount.
OpenAI — GPT-5.2	`cachedInput`	Derived at 10% of input; no residency uplift.
OpenAI — GPT-5.2	`batchInput`	Derived at 50% of input.
OpenAI — GPT-5.2	`batchOutput`	Derived at 50% of output.
OpenAI — GPT-5	`cachedInput`	Derived at 10% of input.
OpenAI — GPT-5	`batchInput`	Derived at 50% of input.
OpenAI — GPT-5	`batchOutput`	Derived at 50% of output.
OpenAI — GPT-5.5 Pro	`cachedInput`	Derived at 10% of input — OpenAI does not publish a cached rate for *-pro models; using the family convention.
OpenAI — GPT-5.5 Pro	`batchInput`	Derived at 50% of input.
OpenAI — GPT-5.5 Pro	`batchOutput`	Derived at 50% of output.
OpenAI — GPT-5.2 Pro	`cachedInput`	Derived at 10% of input — pro-tier convention.
OpenAI — GPT-5.2 Pro	`batchInput`	Derived at 50% of input.
OpenAI — GPT-5.2 Pro	`batchOutput`	Derived at 50% of output.
OpenAI — GPT-5.1	`batchInput`	Derived at 50% of input.
OpenAI — GPT-5.1	`batchOutput`	Derived at 50% of output.
OpenAI — GPT-5 Pro	`batchInput`	Derived at 50% of input.
OpenAI — GPT-5 Pro	`batchOutput`	Derived at 50% of output.
OpenAI — GPT-5 Nano	`cachedInput`	Derived at 10% of input.
OpenAI — GPT-5 Nano	`batchInput`	Derived at 50% of input.
OpenAI — GPT-5 Nano	`batchOutput`	Derived at 50% of output.
Google — Gemini 3 Flash	`cachedInput`	Derived at 10% of input — Google caching discount convention ~90%.
Google — Gemini 3.1 Flash-Lite	`cachedInput`	Derived at 10% of input — Google caching convention.
Google — Gemini 3.1 Flash-Lite	`batchInput`	Derived at 50% of input — Google Batch API uniform 50% discount.
Google — Gemini 3.1 Flash-Lite	`batchOutput`	Derived at 50% of output — Google Batch API uniform 50% discount.
Google — Gemini 2.5 Pro	`cachedInput`	Derived at 10% of input.
Google — Gemini 2.5 Flash	`cachedInput`	Derived at 10% of input.
Google — Gemini 2.5 Flash-Lite	`cachedInput`	Derived at 10% of input — Google caching convention.
Google — Gemini 2.5 Flash-Lite	`batchInput`	Derived at 50% of input — Google Batch API uniform 50% discount.
Google — Gemini 2.5 Flash-Lite	`batchOutput`	Derived at 50% of output — Google Batch API uniform 50% discount.
xAI — Grok 4 (legacy)	`cachedInput`	Extrapolated at 25% of base.

Vendor / Model

Field

Why it’s inferred

Anthropic — Claude Sonnet 4.6

cachedInput

Derived at 10% of input rate — Anthropic publishes 90% cache-hit discount on this tier.

Anthropic — Claude Sonnet 4.5

cachedInput

Derived at 10% of input rate; same 90% cache-hit convention as Sonnet 4.6.

Anthropic — Claude Sonnet 4.5

batchInput

Derived at 50% of standard input — Anthropic documents uniform 50% Batch discount.

Anthropic — Claude Sonnet 4.5

batchOutput

Derived at 50% of standard output — Anthropic documents uniform 50% Batch discount.

Anthropic — Claude Haiku 4.5

cachedInput

Derived at 10% of input rate — Anthropic 90% cache-hit discount convention.

OpenAI — GPT-5.4 Mini

cachedInput

Derived at 10% of input — OpenAI documents automatic 90% discount on cache hits across GPT-5.x tier.

OpenAI — GPT-5.4 Nano

cachedInput

Derived at 10% of input — OpenAI 90% cache-hit convention.

OpenAI — GPT-5.4 Nano

batchInput

Derived at 50% of input — OpenAI Batch API uniform 50% discount.

OpenAI — GPT-5.4 Nano

batchOutput

Derived at 50% of output — OpenAI Batch API uniform 50% discount.

OpenAI — GPT-5.4 Pro

cachedInput

Derived at 10% of input — OpenAI 90% cache-hit convention.

OpenAI — GPT-5.4 Pro

batchInput

Derived at 50% of input — OpenAI Batch API uniform 50% discount.

OpenAI — GPT-5.4 Pro

batchOutput

Derived at 50% of output — OpenAI Batch API uniform 50% discount.

OpenAI — GPT-5.2

cachedInput

Derived at 10% of input; no residency uplift.

OpenAI — GPT-5.2

batchInput

Derived at 50% of input.

OpenAI — GPT-5.2

batchOutput

Derived at 50% of output.

OpenAI — GPT-5

cachedInput

Derived at 10% of input.

OpenAI — GPT-5

batchInput

Derived at 50% of input.

OpenAI — GPT-5

batchOutput

Derived at 50% of output.

OpenAI — GPT-5.5 Pro

cachedInput

Derived at 10% of input — OpenAI does not publish a cached rate for *-pro models; using the family convention.

OpenAI — GPT-5.5 Pro

batchInput

Derived at 50% of input.

OpenAI — GPT-5.5 Pro

batchOutput

Derived at 50% of output.

OpenAI — GPT-5.2 Pro

cachedInput

Derived at 10% of input — pro-tier convention.

OpenAI — GPT-5.2 Pro

batchInput

Derived at 50% of input.

OpenAI — GPT-5.2 Pro

batchOutput

Derived at 50% of output.

OpenAI — GPT-5.1

batchInput

Derived at 50% of input.

OpenAI — GPT-5.1

batchOutput

Derived at 50% of output.

OpenAI — GPT-5 Pro

batchInput

Derived at 50% of input.

OpenAI — GPT-5 Pro

batchOutput

Derived at 50% of output.

OpenAI — GPT-5 Nano

cachedInput

Derived at 10% of input.

OpenAI — GPT-5 Nano

batchInput

Derived at 50% of input.

OpenAI — GPT-5 Nano

batchOutput

Derived at 50% of output.

Google — Gemini 3 Flash

cachedInput

Derived at 10% of input — Google caching discount convention ~90%.

Google — Gemini 3.1 Flash-Lite

cachedInput

Derived at 10% of input — Google caching convention.

Google — Gemini 3.1 Flash-Lite

batchInput

Derived at 50% of input — Google Batch API uniform 50% discount.

Google — Gemini 3.1 Flash-Lite

batchOutput

Derived at 50% of output — Google Batch API uniform 50% discount.

Google — Gemini 2.5 Pro

cachedInput

Derived at 10% of input.

Google — Gemini 2.5 Flash

cachedInput

Derived at 10% of input.

Google — Gemini 2.5 Flash-Lite

cachedInput

Derived at 10% of input — Google caching convention.

Google — Gemini 2.5 Flash-Lite

batchInput

Derived at 50% of input — Google Batch API uniform 50% discount.

Google — Gemini 2.5 Flash-Lite

batchOutput

Derived at 50% of output — Google Batch API uniform 50% discount.

xAI — Grok 4 (legacy)

cachedInput

Extrapolated at 25% of base.

Pinecone vs pgvector vs Qdrant vs 5 more

Vector DB Cost Calculator

Results

Go deeper

Need help using this calculator for your workloads?

Methodology

Primary sources

Inferred values (marked with * in calculator tables)

Immediate steps you can take