A100 80GB SXM4 vs V100 32GB SXM2 for Local LLMs

The A100 80GB SXM4 has about 127% more memory bandwidth (2,039 vs 900 GB/s). Batch-1 decode tracks bandwidth closely, so a model that fits both cards should decode roughly 127% faster on the A100 80GB SXM4. The A100 80GB SXM4 fits more of the catalog at Q4/8K — 55 vs 43 of 73 models (80 vs 32 GB VRAM). Largest model that fits the A100 80GB SXM4 but not the V100 32GB SXM2: Mistral-Medium-3.5-128B (128.0B). MSRP is $15,000 for the A100 80GB SXM4 vs $10,000 for the V100 32GB SXM2. MSRP is a launch list price, not live retail — Amazon prices move; click through for the current price. Not retail-shoppable: A100 80GB SXM4, V100 32GB SXM2 (datacenter-class; cloud rental is the realistic way to use it). In the Q4/8K head-to-head below (57 shared models), the A100 80GB SXM4 posts the higher tok/s in 57 rows vs 0. Methodology.

Spec comparison

	A100 80GB SXM4	V100 32GB SXM2
Vendor	NVIDIA	NVIDIA
VRAM	80 GB	32 GB
Memory type	HBM2e	HBM2
Memory bandwidth	2,039 GB/s	900 GB/s
TDP	400 W	300 W
MSRP (list)	$15,000	$10,000
Buy	—	—

MSRP is the launch list price, not live retail. Buy links are Amazon affiliate links — the price after you click is the live price.

Q4 / 8K head-to-head — top 15 of 57 shared models

Model	Params (B)	A100 80GB SXM4 tok/s	A100 80GB SXM4 fits	V100 32GB SXM2 tok/s	V100 32GB SXM2 fits	Winner
MiniCPM5-1B	1.08	2049.0 est.	Yes	895.6 est.	Yes	A100 80GB SXM4
Qwen3.5-2B	2.0	1106.4 est.	Yes	483.6 est.	Yes	A100 80GB SXM4
DeepSeek-R1-0528-Qwen3-8B-layer-mix-bpw-3.8-mlx	2.3	962.1 est.	Yes	420.5 est.	Yes	A100 80GB SXM4
Qwen3.6-35B-A3B	36.0	737.6 est.	Yes	322.4 est.	Yes	A100 80GB SXM4
Qwen3.5-35B-A3B	34.7	737.6 est.	Yes	322.4 est.	Yes	A100 80GB SXM4
GLM-4.7-Flash	31.2	737.6 est.	Yes	322.4 est.	Yes	A100 80GB SXM4
Qwen3-30B-A3B-Instruct-2507	30.5	737.6 est.	Yes	322.4 est.	Yes	A100 80GB SXM4
Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16	30.0	737.6 est.	Yes	322.4 est.	Yes	A100 80GB SXM4
Qwen3-Coder-Next	79.7	737.6 est.	Yes	268.3 est.	No	A100 80GB SXM4
Ornith-1.0-35B	35.0	737.6 est.	Yes	322.4 est.	Yes	A100 80GB SXM4
Qwen3-Next-80B-A3B-Instruct	81.3	737.6 est.	Yes	258.2 est.	No	A100 80GB SXM4
qwen35b-a3b-fable-sft-abliterated	34.7	737.6 est.	Yes	322.4 est.	Yes	A100 80GB SXM4
Llama-3.2-3B-Instruct	3.2	691.5 est.	Yes	302.3 est.	Yes	A100 80GB SXM4
gpt-oss-20b	21.5	614.7 est.	Yes	268.7 est.	Yes	A100 80GB SXM4
gemma-4-26B-A4B-it	25.2	582.3 est.	Yes	254.5 est.	Yes	A100 80GB SXM4

est. = bandwidth-based estimate; measured = curated real benchmark that overrides the estimate. Fits = weights + KV(8K) + 1 GB overhead ≤ 95% of VRAM.

Fit coverage at Q4 / 8K

	A100 80GB SXM4	V100 32GB SXM2
Catalog models that fit	55 of 73	43 of 73
Largest exclusive fit	Mistral-Medium-3.5-128B (128.0B)	—

A100 80GB SXM4 details → · V100 32GB SXM2 details →