Skip to main content
aifinhub

Engine-computed reference · 6×5 grid · 30 cells

Batch vs Realtime Savings Reference Grid

The monthly saving from running an overnight batch tier instead of realtime inference, for every combination of daily job volume and output size across this grid, priced on Claude Sonnet 4.6. Each cell is a live run of the Batch vs Realtime Cost Calculator engine; no value on this page was entered by hand.

When a workload can tolerate a 24-hour deadline, the batch tier applies a flat discount, so the absolute monthly saving grows with both job volume and output size. Across this grid the saving runs from $270/month (lightest workload at 1,000 jobs/day, 200 output tokens) up to $112,500/month (heaviest at 100,000 jobs/day, 4000 output tokens). The discount is a constant 50% off the realtime daily cost, and at the 24h deadline every cell on this grid is batch-eligible. For when overnight batching pays off, see batch vs realtime overnight cost. Education only — not investment advice.

Monthly batch savings by jobs per day × output tokens

Rows = jobs per day. Columns = output tokens per job. Each value is the engine's savingsPerMonth output in USD.

Monthly batch savings for each daily job volume and output size.
jobs/day \ out tok 2005001,0002,0004,000
1,000 $270 $338 $450 $675 $1,125
5,000 $1,350 $1,688 $2,250 $3,375 $5,625
10,000 $2,700 $3,375 $4,500 $6,750 $11,250
25,000 $6,750 $8,438 $11,250 $16,875 $28,125
50,000 $13,500 $16,875 $22,500 $33,750 $56,250
100,000 $27,000 $33,750 $45,000 $67,500 $112,500

Headline metric: monthly batch savings (USD). The CSV download below also carries the realtime daily cost, batch daily cost, and batch eligibility per cell.

Download CSV (30 rows)

Provenance

Engine
Batch vs Realtime Cost Calculator (batch-vs-realtime-cost-calculator) — computed live from /engines/batch-vs-realtime-cost-calculator.js
Grid
jobs/day ∈ {1,000, 5,000, 10,000, 25,000, 50,000, 100,000} × output tokens ∈ {200, 500, 1000, 2000, 4000} = 30 cells
Fixed inputs
model=claude-sonnet-4-6, input_tokens_per_job=5000, deadline_hours=24
Computed
2026-05-23, recomputed in CI on every build

The engine is deterministic: token accounting and the batch discount are closed-form, so the same input always returns the same output. The full method — the token cost per job, the batch SLA gate, and the discount rate — is documented at the Batch vs Realtime Cost Calculator methodology page. For the broader economics, see batch API economics for finance.

Reference grids are planning aids, not financial, tax, or investment advice. Prices reflect the engine's published rate card and may differ from your billing.

Planning estimates only — not financial, tax, or investment advice.