← AI Hashrate

A100 80GB SXM4 vs Gaudi 3 128GB for Local LLMs

The Gaudi 3 128GB has about 61% more memory bandwidth (3,276 vs 2,039 GB/s). Batch-1 decode tracks bandwidth closely, so a model that fits both cards should decode roughly 61% faster on the Gaudi 3 128GB. The Gaudi 3 128GB fits more of the catalog at Q4/8K — 47 vs 46 of 61 models (128 vs 80 GB VRAM). Largest model that fits the Gaudi 3 128GB but not the A100 80GB SXM4: DeepSeek-V4-Flash (158.1B). MSRP is $15,000 for the A100 80GB SXM4 vs $12,500 for the Gaudi 3 128GB. MSRP is a launch list price, not live retail — Amazon prices move; click through for the current price. Not retail-shoppable: A100 80GB SXM4, Gaudi 3 128GB (datacenter-class; cloud rental is the realistic way to use it). In the Q4/8K head-to-head below (54 shared models), the Gaudi 3 128GB posts the higher tok/s in 53 rows vs 1. Methodology.

Spec comparison

A100 80GB SXM4Gaudi 3 128GB
VendorNVIDIAIntel
VRAM80 GB128 GB
Memory typeHBM2eHBM2e
Memory bandwidth2,039 GB/s3,276 GB/s
TDP400 W600 W
MSRP (list)$15,000$12,500
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 54 shared models

ModelParams (B)A100 80GB SXM4 tok/sA100 80GB SXM4 fitsGaudi 3 128GB tok/sGaudi 3 128GB fitsWinner
Qwen3-0.6B0.81621.9 est.Yes2605.9 est.YesGaudi 3 128GB
gemma-3-1b-it1.01297.5 est.Yes2084.7 est.YesGaudi 3 128GB
Llama-3.2-1B-Instruct1.21081.3 est.Yes1737.3 est.YesGaudi 3 128GB
Qwen3-1.7B2.0648.8 est.Yes1042.4 est.YesGaudi 3 128GB
Qwen3-30B-A3B30.5432.5 est.Yes694.9 est.YesGaudi 3 128GB
Qwen3-Coder-30B-A3B-Instruct30.5432.5 est.Yes694.9 est.YesGaudi 3 128GB
GLM-4.7-Flash31.2432.5 est.Yes694.9 est.YesGaudi 3 128GB
Qwen3-30B-A3B-Instruct-250730.5432.5 est.Yes694.9 est.YesGaudi 3 128GB
Qwen3-Next-80B-A3B-Instruct81.3432.5 est.Yes694.9 est.YesGaudi 3 128GB
Qwen3-Coder-Next79.7432.5 est.Yes694.9 est.YesGaudi 3 128GB
Llama-3.2-3B-Instruct3.2405.5 est.Yes651.5 est.YesGaudi 3 128GB
MiniMax-M2.7228.7143.0 est.No588.4 est.NoGaudi 3 128GB
gpt-oss-20b21.5360.4 est.Yes579.1 est.YesGaudi 3 128GB
Phi-4-mini-instruct3.8341.5 est.Yes548.6 est.YesGaudi 3 128GB
Phi-4-mini-reasoning3.8341.5 est.Yes548.6 est.YesGaudi 3 128GB

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 SXM4Gaudi 3 128GB
Catalog models that fit46 of 6147 of 61
Largest exclusive fit—DeepSeek-V4-Flash (158.1B)

A100 80GB SXM4 details → · Gaudi 3 128GB details →