B200 192GB SXM vs Tesla P4 8GB for Local LLMs

The B200 192GB SXM has about 4067% more memory bandwidth (8,000 vs 192 GB/s). Batch-1 decode tracks bandwidth closely, so a model that fits both cards should decode roughly 4067% faster on the B200 192GB SXM. The B200 192GB SXM fits more of the catalog at Q4/8K — 57 vs 16 of 69 models (192 vs 8 GB VRAM). Largest model that fits the B200 192GB SXM but not the Tesla P4 8GB: MiMo-V2.5 (310.0B). MSRP is $40,000 for the B200 192GB SXM vs $80 for the Tesla P4 8GB. MSRP is a launch list price, not live retail — Amazon prices move; click through for the current price. Not retail-shoppable: B200 192GB SXM (datacenter-class; cloud rental is the realistic way to use it). In the Q4/8K head-to-head below (39 shared models), the B200 192GB SXM posts the higher tok/s in 39 rows vs 0. Methodology.

Spec comparison

	B200 192GB SXM	Tesla P4 8GB
Vendor	NVIDIA	NVIDIA
VRAM	192 GB	8 GB
Memory type	HBM3e	GDDR5
Memory bandwidth	8,000 GB/s	192 GB/s
TDP	1000 W	75 W
MSRP (list)	$40,000	$80
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 39 shared models

Model	Params (B)	B200 192GB SXM tok/s	B200 192GB SXM fits	Tesla P4 8GB tok/s	Tesla P4 8GB fits	Winner
MiniCPM5-1B	1.08	4713.8 est.	Yes	113.1 est.	Yes	B200 192GB SXM
Qwen3.5-2B	2.0	2545.5 est.	Yes	61.1 est.	Yes	B200 192GB SXM
DeepSeek-R1-0528-Qwen3-8B-layer-mix-bpw-3.8-mlx	2.3	2213.4 est.	Yes	53.1 est.	Yes	B200 192GB SXM
Qwen3.6-35B-A3B	36.0	1697.0 est.	Yes	17.0 est.	No	B200 192GB SXM
Qwen3.5-35B-A3B	34.7	1697.0 est.	Yes	18.2 est.	No	B200 192GB SXM
GLM-4.7-Flash	31.2	1697.0 est.	Yes	21.0 est.	No	B200 192GB SXM
Qwen3-30B-A3B-Instruct-2507	30.5	1697.0 est.	Yes	21.9 est.	No	B200 192GB SXM
Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16	30.0	1697.0 est.	Yes	24.1 est.	No	B200 192GB SXM
Ornith-1.0-35B	35.0	1697.0 est.	Yes	17.0 est.	No	B200 192GB SXM
qwen35b-a3b-fable-sft-abliterated	34.7	1697.0 est.	Yes	17.3 est.	No	B200 192GB SXM
Llama-3.2-3B-Instruct	3.2	1590.9 est.	Yes	38.2 est.	Yes	B200 192GB SXM
gpt-oss-20b	21.5	1414.1 est.	Yes	33.6 est.	No	B200 192GB SXM
gemma-4-26B-A4B-it	25.2	1339.7 est.	Yes	23.9 est.	No	B200 192GB SXM
Qwen3-4B-Instruct-2507	4.0	1272.7 est.	Yes	30.5 est.	Yes	B200 192GB SXM
NVIDIA-Nemotron-3-Nano-4B-BF16	4.0	1272.7 est.	Yes	30.5 est.	Yes	B200 192GB SXM

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

	B200 192GB SXM	Tesla P4 8GB
Catalog models that fit	57 of 69	16 of 69
Largest exclusive fit	MiMo-V2.5 (310.0B)	—

B200 192GB SXM details → · Tesla P4 8GB details →