← AI Hashrate

H200 141GB SXM5 vs RX 9070 16GB for Local LLMs

The H200 141GB SXM5 has about 650% more memory bandwidth (4,800 vs 640 GB/s). Batch-1 decode tracks bandwidth closely, so a model that fits both cards should decode roughly 650% faster on the H200 141GB SXM5. The H200 141GB SXM5 fits more of the catalog at Q4/8K — 48 vs 25 of 61 models (141 vs 16 GB VRAM). Largest model that fits the H200 141GB SXM5 but not the RX 9070 16GB: MiniMax-M2.7 (228.7B). MSRP is $35,000 for the H200 141GB SXM5 vs $549 for the RX 9070 16GB. MSRP is a launch list price, not live retail — Amazon prices move; click through for the current price. Not retail-shoppable: H200 141GB SXM5 (datacenter-class; cloud rental is the realistic way to use it). In the Q4/8K head-to-head below (44 shared models), the H200 141GB SXM5 posts the higher tok/s in 44 rows vs 0. Methodology.

Spec comparison

H200 141GB SXM5RX 9070 16GB
VendorNVIDIAAMD
VRAM141 GB16 GB
Memory typeHBM3eGDDR6
Memory bandwidth4,800 GB/s640 GB/s
TDP700 W220 W
MSRP (list)$35,000$549
Buy—Search Amazon (affiliate)

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 44 shared models

ModelParams (B)H200 141GB SXM5 tok/sH200 141GB SXM5 fitsRX 9070 16GB tok/sRX 9070 16GB fitsWinner
Qwen3-0.6B0.83818.2 est.Yes509.1 est.YesH200 141GB SXM5
gemma-3-1b-it1.03054.5 est.Yes407.3 est.YesH200 141GB SXM5
Llama-3.2-1B-Instruct1.22545.5 est.Yes339.4 est.YesH200 141GB SXM5
Qwen3-1.7B2.01527.3 est.Yes203.6 est.YesH200 141GB SXM5
Qwen3-30B-A3B30.51018.2 est.Yes293.9 est.NoH200 141GB SXM5
Qwen3-Coder-30B-A3B-Instruct30.51018.2 est.Yes293.9 est.NoH200 141GB SXM5
GLM-4.7-Flash31.21018.2 est.Yes282.1 est.NoH200 141GB SXM5
Qwen3-30B-A3B-Instruct-250730.51018.2 est.Yes293.9 est.NoH200 141GB SXM5
Qwen3-Next-80B-A3B-Instruct81.31018.2 est.Yes46.3 est.NoH200 141GB SXM5
Qwen3-Coder-Next79.71018.2 est.Yes48.1 est.NoH200 141GB SXM5
Llama-3.2-3B-Instruct3.2954.5 est.Yes127.3 est.YesH200 141GB SXM5
gpt-oss-20b21.5848.5 est.Yes113.1 est.YesH200 141GB SXM5
Phi-4-mini-instruct3.8803.8 est.Yes107.2 est.YesH200 141GB SXM5
Phi-4-mini-reasoning3.8803.8 est.Yes107.2 est.YesH200 141GB SXM5
gemma-3-4b-it4.3783.2 est.Yes104.4 est.YesH200 141GB SXM5

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

H200 141GB SXM5RX 9070 16GB
Catalog models that fit48 of 6125 of 61
Largest exclusive fitMiniMax-M2.7 (228.7B)—

H200 141GB SXM5 details → · RX 9070 16GB details →