A100 vs B300
A100 starts at $0.89/GPU-hour on-demand; B300 starts at $4.99/GPU-hour on-demand. Compare the exact variants, memory and published specifications below. Peak compute figures do not establish application speed.
A100
B300
Spec sheet
| A100 | B300 | |
|---|---|---|
| Spec variant | A100 SXM 40GB | B300 |
| Architecture | Ampere | Blackwell Ultra |
| VRAM (spec variant) | 40 GB | 288 GB |
| FP16 TFLOPS (dense) | 312 | 2250 |
| Memory bandwidth | 1555 GB/s | 8000 GB/s |
| TDP | 400 W | 1400 W |
| Interconnect | NVLink 600GB/s | NVLink 1800GB/s |
| Launched | 2020 | 2025 |
Rental price per manufacturer-rated FP16 compute: A100 $2.85 per PFLOP-hour vs B300 $2.22. Formula: cheapest $/hr ÷ dense FP16 PFLOPS (1 PFLOP = 1000 TFLOPS). This uses A100 SXM 40GB and B300, whose datasheets may differ from the variants above. It is not measured job throughput. Specs: A100 manufacturer, B300 manufacturer.
What they cost to rent (2026-10-02)
| A100 | B300 | |
|---|---|---|
| Cheapest on-demand $/hr | $0.89 | $4.99 |
| Cheapest spot/community $/hr | $0.679 | $4.99 |
| Cheapest source | Jarvislabs | Prime Intellect |
| Providers offering it | 21 | 13 |
Lowest recorded on-demand rate per configuration; highlighted = cheaper. Spot/marketplace tiers shown on the individual pages.
FAQ
A100 is cheaper: the A100 rents from $0.89 per GPU-hour and the B300 from $4.99 on-demand, updated 2026-10-02.
The largest recorded A100 variant has 80 GB and the largest B300 variant has 288 GB. Other variants may have less memory. The B300 has more.
Compare more rental options: the screener · price movers · methodology.