Pricing
Powerful GPUs at a fraction of the cost.
Over 2x cheaper than the competition.
Cloud GPUs
Run inference on datacenter GPUs. Billed per second of compute, with no cold start fees.
| Acceleration | |
|---|---|
CPU remote_cpu | $0.10 / hr |
Nvidia A10G remote_a10 | $0.55 / hr |
Nvidia L40S remote_l40s | $0.50 / hr |
Nvidia A100 remote_a100 | $1.25 / hr |
Nvidia H100 remote_h100 | $1.99 / hr |
Nvidia B200 remote_b200 | $3.99 / hr |
AMD MI350X remote_mi350x | — |
AMD MI355X remote_mi355x | — |
Qualcomm Cloud AI 100 remote_qaic100 | — |
On-device Inference
Deploy models to run directly on your users' devices. Pay a one-time cost to deploy a model to a device, then run infinitely many predictions at zero cost.
| Tier | Price |
|---|---|
| First 100 model deployments | Free |
| Additional model deployments | $0.10 / device |
Plans
Start for free, then scale with your team. Every plan includes pay-as-you-go cloud compute and on-device deployments.
Developer
$0+ usage
Start building and deploying AI models in your apps.
What's included
- Up to 5 compiled models.
- Cloud inference on all GPU tiers.
- 100 free on-device deployments.
- Community support.
Team
$200+ usage
Go from prototype to production with your team.
What's included
- 4 seats + $50/seat/mo.
- Unlimited compiled models.
- Usage analytics and benchmarks.
- Priority support.
Enterprise
Custom
Deploy Muna across your enterprise, to internal teams and customers.
What's included
- On-prem & VPC deployments.
- Dedicated GPU capacity.
- Custom model & hardware compilation.
- SLA with uptime guarantee.
Frequently Asked Questions
Answering a few common questions.