Question 1

What AI models can run on H200 SXM?

Accepted Answer

The H200 SXM (141 GB VRAM) can run models including Llama-3.2-1B-Instruct-SpinQuant_INT4_EO8, Llama-3.2-1B-Instruct-QLORA_INT4_EO8, Llama-3.2-3B-Instruct-QLORA_INT4_EO8, Llama-2-7b, Llama-3.2-3B-Instruct-SpinQuant_INT4_EO8 and more — 10 compatible models indexed so far. Compatible models are determined by whether the model's memory footprint fits within the GPU's 141 GB VRAM, potentially using quantization (Q8, Q6, or Q4) to reduce memory usage.

Question 2

What is the cheapest cloud option for H200 SXM?

Accepted Answer

The cheapest on-demand option for the H200 SXM is Vast.ai at $2.35/hr (1x-H200-32-65536). Spot/preemptible pricing is available at $0.00/hr. We track pricing across 4 providers to help you find the best deal.

Question 3

Is H200 SXM good for fine-tuning?

Accepted Answer

Yes, the H200 SXM is excellent for fine-tuning. With 141 GB of VRAM, it can handle full fine-tuning of models up to ~30B parameters and LoRA/QLoRA fine-tuning of much larger models. It's a top-tier choice for training workloads.

Question 4

How much VRAM does H200 SXM have?

Accepted Answer

The H200 SXM has 141 GB of VRAM (HBM3e). It also offers 4800 GB/s memory bandwidth, 989.5 TFLOPS of FP16 compute, FP8 support for efficient inference.

Question 5

What quantization formats work on H200 SXM?

Accepted Answer

The H200 SXM supports the following quantization formats: FP16 (full precision), FP8, Q8 (8-bit), Q6 (6-bit), Q4 (4-bit). FP16 works for models up to ~40B parameters. FP8 halves memory vs FP16 with minimal quality loss. Q4 quantization lets models up to ~479B parameters (by VRAM estimate) fit on the H200 SXM.

Provider	On-demand	Spot	Actions
Vast.ai	$2.350/hr	$0.000/hr	Rent
Nebius	$4.500/hr	$2.450/hr	—

H200 SXM

Specifications

Cloud pricing

Compatible models

Compare this GPU

Frequently asked questions

VRAM	141 GB
Memory type	HBM3e
Bandwidth	4800 GB/s
FP16 TFLOPS	989.5
FP8 support	Yes