COMPARE / A10 VS A30

A10 vs A30

Specs from the chip makers' datasheets; prices from the live Compute Cafe panel, refreshed hourly. Every rate is an on-demand published price with a source receipt.

CHEAPEST A10
$0.150/hr
SpotGPUs · median $1.56 · 11 providers
CHEAPEST A30
$0.080/hr
SpotGPUs · median $0.74 · 7 providers
CHEAPER RIGHT NOW
A30
live panel, 2026-09-27
A10A30
MAKERNVIDIANVIDIA
ARCHITECTUREAmpereAmpere
LAUNCHED20212021
MEMORY24 GB GDDR624 GB HBM2
MEMORY BANDWIDTH600 GB/s933 GB/s
TENSOR COMPUTE125 TFLOPS FP16 tensor dense (250 sparse)165 TFLOPS FP16 tensor dense (330 sparse)
POWER DRAW150 W165 W
INTERCONNECTPCIe 4.0 x16PCIe 4.0; NVLink bridge 200 GB/s

HOW TO CHOOSE

A30 vs A10
Same memory size, 55% more bandwidth from HBM2, slightly more compute. For LLM serving the A30 is usually the better chip per dollar.
Workload fit
A10: mid-size inference, virtual workstations. A30: bandwidth-sensitive inference, multi-tenant serving.

A10 VS A30 - FAQ

Which is cheaper to rent, A10 or A30?
By today's published on-demand rates: A10 starts at $0.150/GPU/hr (median $1.56, 11 providers) and A30 starts at $0.080/GPU/hr (median $0.74, 7 providers). The cheapest live listing is the A30 at SpotGPUs. Figures recompute from the panel on every refresh.
A10 or A30 - which should I rent?
A10: The A10 fills Ampere's single-slot inference slot: 24 GB, respectable FP16 throughput, 150 W. A30: The A30 is the quiet value part of the Ampere line: HBM2 bandwidth (933 GB/s) at 165 W, with MIG support for slicing into four isolated instances. The side-by-side spec table and live price rows above give the full picture.
Where can I rent A10 and A30?
11 providers publish A10 rates and 7 publish A30 rates on Compute Cafe. Every row links to the provider's own published price - committed-use and reserved rates are excluded.

Full pricing: A10 · A30 · all comparisons