> ## Documentation Index
> Fetch the complete documentation index at: https://docs.qbraid.com/llms.txt
> Use this file to discover all available pages before exploring further.
# On-Demand Pricing
## Credits
Access to both classical and quantum computing resources is managed through the qBraid credits system, which operates on a pay-as-you-go basis.
Each qBraid credit is worth \$0.01 USD, so a quantum job costing \$3.80 would subtract 380 credits from your qBraid balance. Credits can be
[purchased](https://account.qbraid.com/account/wallet) from your account page, or [redeemed](/v2/account/wallet/buy-credits#apply-a-promo-code-on-the-wallet)
using a promo code. You can check your current credit balance on your [account page](https://account.qbraid.com/account/wallet), under
the "Profile" drop-down in the top right of [qBraid Lab](/v2/lab/user-guide/overview), or using the [qBraid-CLI](/v2/cli/api-reference/qbraid_account#qbraid-account-credits).
Prices for both classical and quantum computing resources are subject to change.
## Quantum Computers
There are three pricing components for on-demand use of a quantum computer, or quantum processing unit (QPU), on qBraid: a per-minute fee, per-task fee, and per-shot fee.
A shot is a single execution of a quantum algorithm on a QPU. For example, a shot is a single pass through each stage of a complete quantum circuit on a gate-based QPU from
IonQ, Rigetti, IQM, or AQT, or the time evolution of a Hamiltonian on a QPU from QuEra. A task is a sequence of repeated shots based on the same circuit design or Hamiltonian.
You define how many shots you want included in a task when you submit it.
### Pricing Structure
* Per-minute pricing: Applies to specific QPUs (e.g., Pasqal Fresnel) where execution time is billed
* Per-task pricing: Consistent across most QPUs (typically 30 credits)
* Per-shot pricing: Varies by QPU type (\<1 to 8 credits for current offerings)
For gate-based QPUs, neither per-shot nor per-minute prices are affected by the number or type of gates used. Note that error mitigation on IonQ's Aria QPU requires a minimum of 2,500 shots per task.
| Hardware Provider | QPU | Per-minute price | Per-task price | Per-shot price |
| :---------------- | :----------------- | :--------------- | :------------- | :------------- |
| AQT | IBEX Q1 | 0 credits | 30 credits | 2.35 credit |
| QuEra | Aquila | 0 credits | 30 credits | 1 credit |
| IQM | Garnet | 0 credits | 30 credits | 0.145 credits |
| IQM | Emerald | 0 credits | 30 credits | 0.16 credits |
| IonQ | Forte-1 | 0 credits | 30 credits | 8 credits |
| IonQ | Forte-Enterprise-1 | 0 credits | 30 credits | 8 credits |
| Rigetti (AWS) | Cepheus-1-108Q | 0 credits | 30 credits | 0.0425 credits |
| Rigetti (Direct) | Cepheus-1-108Q | 12000 credits | 0 credits | 0 credits |
| Pasqal | Fresnel | 500 credits | 0 credits | 0 credits |
## Quantum Circuit Simulators
Verify your circuits using a simulator before you run it on a QPU, so you can fine-tune your circuit without incurring charges for QPU usage.
Although the results from running the circuit on a simulator may not be identical to the results from running the circuit on a QPU, you can
identify coding errors or configuration issues using a simulator.
| Provider | Name | Per-minute price | Access |
| ---------- | :------------------- | :--------------- | :------ |
| qBraid | QIR Simulator | 0 credits | public |
| AWS | DM1 | 7.5 credits | public |
| AWS | SV1 | 7.5 credits | public |
| AWS | TN1 | 27.5 credits | public |
| Rigetti | QVM Simulator | 0 credits | public |
| IonQ | Simulator | 0 credits | public |
| Quantinuum | H2-1 Syntax checker | 0 credits | public |
| Quantinuum | H2-1 Emulator | 0 credits | public |
| Pasqal | Emulator | 25 credits | public |
| Quobly | Alloy Forge Emulator | 0 credits | limited |
| Equal1 | Bell-1 Simulator | 12.5 credits | limited |
| NEC | Vector Annealer | 0 credits | limited |
qBraid offers twelve on-demand quantum circuit simulators for testing quantum algorithms:
* qBraid QIR Simulator: employs sparse state vector simulation based on [arXiv:2105.01533](https://arxiv.org/abs/2105.01533).
* AWS State Vector 1 (SV1): serves as a general purpose quantum circuit simulator.
* AWS Density Matrix 1 (DM1): enables simulating the effect of noise on your circuits.
* AWS Tensor Network 1 (TN1): specializes in simulation of larger scale quantum systems.
* Rigetti QVM Simulator: Quantum Virtual Machine (QVM) open-source simulator.
* IonQ Simulator: supports "ideal" simulations as well as noise models for harmony (legacy), aria-1, aria-2, forte-1, and forte-enterprise-1.
* NEC Vector Annealer: executes proprietary algorithm for annealing processing on NEC’s vector supercomputer "SX-Aurora TSUBASA“
* Quantinuum H2-1 Syntax Checker: Validates quantum programs against the H2-1 compiler.
* Quantinuum H2-1 Emulator: Emulation tool containing a realistic noise model of the actual System Model H2 hardware.
* Pasqal Emulator: Neutral atom tensor network emulator. Runs on a cluster of DGX nodes with NVIDIA A100 GPUs.
* Equal1 Bell-1 Simulator: State-vector simulator supporting Equal1 Bell-1 QPU noise models; GPU acceleration available
* Quobly Alloy Forge Emulator: Emulator of the Quobly Alloy Pioneer 10 qubit QPU.
All simulators automatically scale resources to ensure high-performance. Billing is based on a fixed per-task cost plus the duration
of each simulation, charged per minute in one-millisecond increments, with a minimum billing duration of three seconds per simulation.
The QPU and simulator lists above are subject to change and may be incomplete
or include options that are no longer available. For the most complete and
up-to-date list of quantum devices, visit your [account
page](https://account.qbraid.com/devices).
## CPUs and GPUs
qBraid Lab offers two ways to access compute: **subscription instances** billed against your monthly CPU hours, and **on-demand instances** billed per minute in credits.
### Subscription Instances
Your subscription plan includes a monthly allocation of [CPU hours](/v2/lab/user-guide/usage-limits#compute-hours). The following CPU instances draw from that quota:
| Instance | vCPU | RAM | Credits/Min (Overdraft) |
| :------- | :--- | :--- | :---------------------- |
| Small | 2 | 4GB | 0.20 |
| Medium | 4 | 12GB | 0.40 |
| Large | 8 | 25GB | 0.80 |
If you exceed your monthly CPU hours, you can [auto-switch to credits](/v2/lab/user-guide/usage-limits#auto-switch-to-credits) — usage then bills at the per-minute credit rates shown above. These rates are consistent with on-demand pricing, so there is no difference in cost between overdraft and on-demand for the same compute. See [usage limits](/v2/lab/user-guide/usage-limits) for more details on compute hour tracking.
### On-Demand Instances
On-demand instances are billed per minute in [credits](/v2/home/pricing#credits), with no subscription commitment required. This includes all GPU instances and additional high-spec CPU configurations. See [On-Demand Instances](/v2/lab/user-guide/on-demand-instances) for the full lifecycle guide.
**Subscription discounts on on-demand pricing:**
| Plan | Discount |
| :------- | :------- |
| Free | — |
| Standard | 10% off |
| Pro | 18% off |
Standard and Pro subscribers receive an automatic discount on all on-demand instances (GPUs and on-demand CPUs). The discount is applied at billing time — the effective rate is shown on your dashboard alongside the list price.
The **Slug** column is the profile identifier used with the [qBraid CLI](/v2/cli/api-reference/qbraid_compute). For example:
```console theme={null}
$ qbraid compute up gpu-h100-sxm
```
#### GPU Instances
See [GPUs](/v2/lab/user-guide/gpus) for more details on launching and using GPU instances.
| Slug | GPU Type | VRAM | Credits/Min |
| :----------------- | :--------------------- | :----- | :---------- |
| `gpu-b200-8x` | 8x NVIDIA B200 | 1440GB | 111.50 |
| `gpu-b200-4x` | 4x NVIDIA B200 | 720GB | 56.58 |
| `gpu-b200-2x` | 2x NVIDIA B200 | 360GB | 28.72 |
| `gpu-b200` | 1x NVIDIA B200 | 180GB | 14.57 |
| `gpu-h200` | 1x NVIDIA H200 | 141GB | 9.15 |
| `gpu-h100-8x` | 8x NVIDIA H100 | 640GB | 66.50 |
| `gpu-h100-4x` | 4x NVIDIA H100 | 320GB | 34.08 |
| `gpu-h100-2x` | 2x NVIDIA H100 | 160GB | 17.47 |
| `gpu-h100-sxm` | 1x NVIDIA H100 SXM | 80GB | 8.95 |
| `gpu-gh200` | 1x NVIDIA GH200 | 96GB | 4.78 |
| `gpu-a100-8x` | 8x NVIDIA A100 | 320GB | 46.50 |
| `gpu-a100-4x` | 4x NVIDIA A100 | 160GB | 16.58 |
| `gpu-a100-2x` | 2x NVIDIA A100 | 80GB | 8.30 |
| `gpu-a100-sxm` | 1x NVIDIA A100 SXM | 80GB | 4.15 |
| `gpu-l40s` | 1x NVIDIA L40S | 48GB | 3.80 |
| `gpu-rtx-5090` | 1x NVIDIA RTX 5090 | 32GB | 2.07 |
| `gpu-a10` | 1x NVIDIA A10 | 24GB | 2.70 |
| `gpu-rtx-4090` | 1x NVIDIA RTX 4090 | 24GB | 1.45 |
| `gpu-rtx-6000-ada` | 1x NVIDIA RTX 6000 Ada | 48GB | 1.53 |
| `gpu-l4` | 1x NVIDIA L4 | 24GB | 0.82 |
#### CPU Instances
On-demand CPU instances offer higher-spec configurations beyond what subscription instances provide.
| Slug | vCPU | RAM | Credits/Min |
| :------------- | ---: | :---- | :---------- |
| `cpu-64v-256g` | 64 | 256GB | 6.40 |
| `cpu-32v-128g` | 32 | 128GB | 3.20 |
| `cpu-8v-32g` | 8 | 32GB | 0.80 |
| `cpu-4v-16g` | 4 | 16GB | 0.40 |
| `cpu-4v-6g` | 4 | 6GB | 0.25 |
Additional configurations may be available. Visit your [account
page](https://account.qbraid.com/dashboard) to see the full list of options.