Your cookie choices

We use necessary cookies to make this site work. With your permission, analytics cookies help us understand usage and marketing cookies support advertising.

You can reject non-essential cookies and still use the site. Change your choices anytime in Cookie settings.

Cookie settingsCookie policy

Inferless review

Updated September 28, 2026 · DeployCue Editorial Team · Facts checked September 26, 2026

We earn a commission when you sign up through links on this site. Partner status can affect where a brand appears in our rankings. Partners with a live offer in your country come first, then every other brand by score.

Our verdict

Inferless suits small teams serving models on Nvidia A100, Nvidia A10 and Nvidia T4 hardware, with GPU hours starting at $0.33 and usage metered per second.

Key facts

Best for
Small model deployments on Nvidia A100, Nvidia A10 and Nvidia T4 hardware, metered per second.
GPU models
Nvidia A100, Nvidia A10, Nvidia T4↗
GPU hour from
$0.33↗
Billing granularity
Per second↗
Free credits
10 hours of free credit↗
Interconnect
AWS PrivateLink private VPC endpoint service↗
Compliance certifications
SOC-2 Type II↗
Legal entity
IDQ Innovation Pvt Ltd↗
Headquarters
India↗
Founded
2022↗
Support channels
Email, Help-centre, Community↗
Pricing model
Hourly billing
Official website
inferless.com

Serverless

Price per GB-second
$0.000092 to $0.00149↗

Facts checked September 26, 2026, 5 sources

Apps and profiles

Elsewhere
LinkedInXtowardsscalininference.community

Pros and cons

Pros

  • Usage is metered per second, so a short demo stops paying the moment the model stops running.
  • New accounts receive 10 hours of free credit, enough to put a real model through a test.
  • Private network access is available through AWS PrivateLink private VPC endpoint service.
  • Certifications on record include SOC 2 Type II.

Cons

  • The card range is limited to Nvidia A100, Nvidia A10 and Nvidia T4, so newer datacenter parts are out of reach.
  • Support runs through email, help centre and community only, with no phone line offered.
  • The only pricing model listed is hourly, so there is no stated reserved rate for steady, long-running jobs.
Inferless

Ready to try Inferless?

Scored 9.0 / 10 by DeployCue.

See alternatives

Who it suits

  • Small teams serving a model on Nvidia A100, Nvidia A10 and Nvidia T4 who want to stop paying once a test ends.
  • Builders working through 10 hours of free credit before moving a real workload onto a paid plan.
  • Regulated teams that need AWS PrivateLink private VPC endpoint service to keep inference traffic inside their own network.

Who should look elsewhere

  • Shoppers after the newest accelerators: ComputePrices.com tracks H100 SXM, A100 SXM and H200, a much wider hardware list.
  • Buyers who want current datacenter parts already listed: SberCloud shows B300, H100 and A100.
  • Teams priced on spare community hardware: Salad runs NVIDIA RTX 3060, NVIDIA RTX 3070 and NVIDIA RTX 3070 Ti rather than datacentre accelerators.

Plans, egress and billing

GPU hours start at $0.33 on the hourly plan, and usage is metered per second, so a short test stops paying once the job ends. New accounts come with 10 hours of free credit to put a model through its paces. That entry rate sits below the €0.63 GPU hour seen across the site, which keeps Inferless competitive for small jobs. Serverless compute runs at $0.000092 to $0.0015 per GB-second.

The company behind it

Inferless is run by IDQ Innovation Pvt Ltd, headquartered in India and founded in 2022. Support runs through email, help centre and community, so help arrives by email, from the help centre or from the community rather than a live call. Teams that keep inference traffic private can connect through AWS PrivateLink private VPC endpoint service, which keeps endpoints on their own network.

How Inferless compares

Frequently asked questions

How much does a GPU hour cost on Inferless?+

GPU hours start at $0.33 on the hourly plan, and usage is metered per second, so you pay only while a model runs. New accounts also get 10 hours of free credit. That entry rate sits below the €0.63 GPU hour seen across the site.

Which GPUs can I deploy on?+

The published card list is Nvidia A100, Nvidia A10 and Nvidia T4. If your model needs the newest datacenter parts, compare H100 SXM, A100 SXM and H200, which tracks a far wider hardware range, and check the hourly rate for the specific card before you commit.

How is usage billed?+

Usage is metered per second on the hourly plan, so a short test run does not carry a full hour of charges. Serverless-style compute is priced per GB-second at $0.000092 to $0.0015, and the platform's listed model is hourly.

Where is Inferless based and who signs the contract?+

Inferless is run by IDQ Innovation Pvt Ltd, headquartered in India and founded in 2022. Buyers who need a named contracting entity on the invoice, or who keep data under a particular jurisdiction, should check that the entity fits their own requirements.

How do I get support?+

Support runs through email, help centre and community, so expect answers by email, from the help centre, or from other users in the community rather than a live call. Teams that need phone escalation should plan for written response times instead.

Can inference traffic stay inside my own network?+

Private connectivity is documented as AWS PrivateLink private VPC endpoint service, which keeps endpoints inside a customer's own network. The platform also lists SOC 2 Type II, so buyers carrying compliance duties can check the certification status before signing anything.

How does this compare with community hardware providers?+

Community providers such as Salad run NVIDIA RTX 3060, NVIDIA RTX 3070 and NVIDIA RTX 3070 Ti, a different hardware mix. Inferless concentrates on Nvidia A100, Nvidia A10 and Nvidia T4 at $0.33 per GPU hour, so the comparison comes down to whether you need a specific card or simply low-cost capacity.

Sources

  • docs.inferless.com/api-reference/aws-privatelink.md checked September 26, 2026

    Backs: Interconnect

    “AWS PrivateLink is an ideal solution for establishing private connectivity between Inferless and your AWS services, your VPC, and on-premises applications”

  • inferless.com checked September 26, 2026

    Backs: Compliance certifications, Headquarters

    “SOC-2 Type II certification”

  • inferless.com/pricing checked September 26, 2026

    Backs: Billing granularity, Free credits, GPU hour from, GPU models, Pricing model, Price per GB-second

    “Pay for what you use: Per-second billing. No upfront costs.”

  • inferless.com/terms checked September 26, 2026

    Backs: Legal entity, Support channels

    “This is an agreement between you or the entity that you represent ... and IDQ Innovation Pvt Ltd ( Hereinafter " IDQPL" OR "Inferless")”

  • baseten.co/blog/announcing-the-acquihire-of-inferless-by-... checked September 26, 2026

    Backs: Founded

    “When we started Inferless in December 2022, our mission was simple”

Compiled from 5 source pages, checked September 26, 2026. We did not open an account.

Not confirmed yet: Egress fee.

Inferless

Inferless

Small model deployments on Nvidia A100, Nvidia A10 and Nvidia T4 hardware, metered per second.

9.0/10See alternatives