Pricing

Pricing Calculator

Estimate your monthly costs with BatchIn and compare against verified platform competitors.

Public pricing source

Loading runtime catalog

Relay order

Waiting on backend state

Payment posture

Stripe default checkout / USDC preview/manual review / x402 preview/dry run

Runtime catalog

Pricing v32.2

Loading model catalog

ModelStatusPublic priceAvailability

Inference

Live catalog

From $0.08 / 1M

The public site only exposes verified inference routes and real lane pricing. Unverified routes do not masquerade as self-serve SKUs.

  • Public price is sourced from the runtime model catalog
  • Default relay order: Gitee AI -> SiliconFlow -> OpenRouter
  • No $0.00 / 1M is shown unless the route is verified free

Batch / Reserved Capacity

Reserved

Quote / Reserved

For batch throughput, reserved token buckets, and operator-assisted supply. Public copy stays capacity-based instead of pretending live stock.

  • Reserved Token Bucket and batch inference are quoted separately
  • Shared, reserved, and special supply are evaluated by region and term
  • Fit for batch processing, training, and enterprise projects

VaaS / Usage Receipts

Plan-enabled

Included with eligible plans

Verification, receipts, and reconciliation around request_id, usage, cost, and audit export. The public site no longer splits this into older workbench SKUs.

  • Usage receipts, audit export, and invoice reconciliation
  • Log retention, cache, and eval controls are packaged through console plans or contract scope
  • Chain anchoring, reserved retention, and enterprise export remain request-access or contract-scoped

Featured routes

Current featured public model lanes

Compute capacity

Quoted by supply and reservation term

The public site does not promise live inventory. Compute is quoted by region, SLA, term length, concurrency, and project scope.

H200 / H100 / H800

Quoted by reservation term, region, and current supply

A800 / A100 / L40S / H20

Fit for inference, training, and mixed reserved pools

RTX 4090 / RTX 5090

Good for smaller labs, agent runtime, and lighter deployments

B200 / B300 / Ascend 910B / 910C

Evaluated case by case for enterprise projects

Managed text routes

Single public price

Public pricing shows a single runtime-verified rate instead of a lowest-price comparison.

  • Public pricing comes from the runtime model catalog rather than a static marketing table
  • Default relay order: Gitee AI -> SiliconFlow -> OpenRouter
  • No $0.00 / 1M is shown unless the route is actually verified free

Pass-through vendor lanes

State-based

Original-vendor relay or transparent pass-through lanes only show public pricing when the backend exposes them as public-ready.

  • Preview and request-access routes are not presented as public self-serve SKUs
  • Pass-through charges still follow upstream billing and contract posture
  • Missing values stay as Contact us or Private preview instead of fabricated numbers