A10 vs L40
A10 starts at $1.00/GPU-hour on-demand; L40 starts at $0.79/GPU-hour on-demand. Compare the exact variants, memory and published specifications below. Peak compute figures do not establish application speed.
A10
L40
Spec sheet
| A10 | L40 | |
|---|---|---|
| Spec variant | A10 | L40 |
| Architecture | Ampere | Ada Lovelace |
| VRAM (spec variant) | 24 GB | 48 GB |
| FP16 TFLOPS (dense) | 125 | 181 |
| Memory bandwidth | 600 GB/s | 864 GB/s |
| TDP | 150 W | 300 W |
| Interconnect | PCIe Gen4 | PCIe Gen4 |
| Launched | 2021 | 2022 |
Rental price per manufacturer-rated FP16 compute: A10 $8.00 per PFLOP-hour vs L40 $4.36. Formula: cheapest $/hr ÷ dense FP16 PFLOPS (1 PFLOP = 1000 TFLOPS). This uses A10 and L40, whose datasheets may differ from the variants above. It is not measured job throughput. Specs: A10 manufacturer, L40 manufacturer.
What they cost to rent (2026-10-02)
| A10 | L40 | |
|---|---|---|
| Cheapest on-demand $/hr | $1.00 | $0.79 |
| Cheapest spot/community $/hr | – | $0.69 |
| Cheapest source | OVHcloud | Thunder Compute |
| Providers offering it | 5 | 6 |
Lowest recorded on-demand rate per configuration; highlighted = cheaper. Spot/marketplace tiers shown on the individual pages.
FAQ
L40 is cheaper: the A10 rents from $1.00 per GPU-hour and the L40 from $0.79 on-demand, updated 2026-10-02.
The largest recorded A10 variant has 24 GB and the largest L40 variant has 48 GB. Other variants may have less memory. The L40 has more.
Compare more rental options: the screener · price movers · methodology.