Pricing
Transparent pricing. 3× more affordable. No email gates.
Per-hour for self-serve. Reserved for enterprise. The numbers below are the numbers you pay.
Pricing model
Pick a path.
The toggle is structural. Don't rely on memory — the page below changes with it.
Pay-as-you-go. Per-second metering. No commitment.
Pricing below is the number you pay — no separate spot, on-demand, or marketplace tiers.
Live pricing
Pay 3× less, today.
Per-second metering, billed hourly. The same NVIDIA hardware your team already uses, with hyperscaler list pricing shown alongside so you can verify against your incumbent.
| SKU | Memory | iFrame $/hr · from / to | AWS | Azure | GCP | You save | |
|---|---|---|---|---|---|---|---|
B300 192 GB HBM3e | 192 GB HBM3e | $3.25 – $4.60 | $12.60 | $12.00 | $12.20 | −73% | Launch |
B200 180 GB HBM3e | 180 GB HBM3e | $3.25 – $4.60 | $10.40 | $9.80 | $10.10 | −67% | Launch |
H200 141 GB HBM3e | 141 GB HBM3e | $2.95 – $4.18 | $8.60 | $8.40 | $8.50 | −65% | Launch |
H100 80GB 80 GB HBM3 | 80 GB HBM3 | $2.25 – $3.18 | $6.98 | $6.85 | $6.90 | −67% | Launch |
- B300192 GB HBM3efrom $3.25/hrup to $4.60 · $12.00 hyperscaler−73% vs hyperscalersLaunch →
- B200180 GB HBM3efrom $3.25/hrup to $4.60 · $9.80 hyperscaler−67% vs hyperscalersLaunch →
- H200141 GB HBM3efrom $2.95/hrup to $4.18 · $8.40 hyperscaler−65% vs hyperscalersLaunch →
- H100 80GB80 GB HBM3from $2.25/hrup to $3.18 · $6.85 hyperscaler−67% vs hyperscalersLaunch →
iFrame range: floor at 5-year reserved + 20% prepay; ceiling at 1-month commit + 10% prepay (1-day terms forced to 100% prepay). Hyperscaler comparison uses each provider's on-demand list pricing for the closest equivalent SKU. Spot, committed-use, and marketplace pricing excluded. Run the calculator.
See full pricingInference (per million tokens)
Full inference pricing →| Model | Input | Output |
|---|---|---|
| Llama 3.1 8B | $0.05 | $0.10 |
| Llama 3.1 70B | $0.32 | $0.55 |
| Mixtral 8×22B | $0.55 | $0.85 |
| DeepSeek V3 | $0.45 | $0.85 |
Predictable workloads, predictable pricing.
Multi-month commitments unlock additional discount, dedicated bare metal, named support, and bundled VPC interconnect setup.
- Reserved capacity (your GPUs, dedicated)
- Named support engineer
- Standard SLA
- Everything in 6 months
- VPC interconnect setup included
- Custom SLA terms
- Everything in 1 year
- BAA, DPA, custom MSA
- Quarterly business reviews
- Audit log streaming
Calculator
What you'd save.#
Move the sliders. Compare against the lowest hyperscaler list price. Share the result as a link.
Compare vs
Side-by-side, honestly.#
vs Hyperscalers
3× more affordable. Instant access. VPC interconnect.
Same NVIDIA hardware as AWS, Azure, and GCP — at one-third the price, available in minutes, peering into your existing VPC.
See the comparisonvs Neoclouds
Lower price. Real VPC. Inference, not just GPUs.
Most neoclouds rent bare GPUs. iFrame is a cloud — managed inference, VPC interconnect, R&D depth, contract flexibility.
See the comparisonDiscount programs
Research and startup credits.
Research
Compute credits for accredited institutions.
Discounted GPU access for principal investigators at accredited universities and research institutions. Grant-friendly procurement supported.
Apply for academic terms →Startups
$25,000 in credits for early-stage AI companies.
Less than $5M raised, less than 50 employees, building something with AI. Approval within one business day.
Apply for credits →- SOC 2 Type II
- HIPAA-ready
- GDPR-ready
- ISO 27001
FAQ
Billing and commitment.#
Pay 3× less today.
Self-serve signup or a reserved-capacity quote — both paths are short.