Skip to content
Updated minutes ago
intel

Gaudi 2

intel · gaudi · 96 GB HBM2e · 600W TDP

VRAM

96 GB

BF16 TFLOPS

432

Bandwidth

2460 GB/s

From

$1.65/hr

Calculate ROI with this GPU →

Spec Sheet

VRAM96 GB HBM2e
Memory Bandwidth2460 GB/s
BF16 TFLOPS432
FP16 TFLOPS432
FP8 TFLOPS432
INT8 TOPS432
TDP600W
InterconnectOAM
Max per Node8
PCIe Gen4
Tensor CoresNo

Pricing by Provider

ProviderOn-DemandReservedSpotBadge
aws$2.30/hr$1.65/hr-Cheapest

Compatible Models (278)

Training Capabilities

Estimated GPU count for full fine-tuning (AdamW, BF16) and QLoRA

Model SizeFull Fine-TuneQLoRA
7B model2 GPUs1 GPU
13B model3 GPUs1 GPU
70B model14 GPUs1 GPU

Energy Efficiency

Estimated tokens/second per Watt for popular models

Mistral 7B
0.56 t/s/WFP8
Qwen 2.5 7B
0.54 t/s/WFP8
Llama 3.1 8B
0.51 t/s/WFP8
Llama 3.1 70B
0.06 t/s/WFP8
Qwen 2.5 72B
0.06 t/s/WFP8

Similar GPUs

GPUVRAMBF16 TFLOPSBW (GB/s)From
Gaudi 3128 GB18353700$2.30/hr
Gaudi 3 HL-325L128 GB18353700$2.65/hr
H2096 GB1484000$0.99/hr
GH20096 GB9904000$2.99/hr
Trainium296 GB7563200$1.95/hr

Methodology Note

Performance estimates for the Gaudi 2are based on InferenceBench's roofline performance model with CUDA kernel-level optimization including FlashAttention v2 and PagedAttention. Memory calculations account for model weights (96 GB HBM2e available), KV-cache allocation, and activation memory. Throughput predictions use the Gaudi 2's rated 2460 GB/s memory bandwidth and 432 BF16 TFLOPS compute capacity as roofline ceilings, with empirical correction factors per GPU architecture (gaudi). See our full methodology.

Frequently Asked Questions

How many AI models can run on Gaudi 2?

The Gaudi 2 can run 278 AI models from our database within a single node. Compatible models range across various parameter sizes depending on the quantization precision (BF16, FP8, INT4). Smaller models fit on a single GPU while larger models may require multi-GPU setups up to 8x Gaudi 2.

What is the Gaudi 2 inference throughput?

The Gaudi 2 delivers 432 BF16 TFLOPS and 432 FP8 TFLOPS with 2460 GB/s memory bandwidth. Actual inference throughput (tokens/sec) depends on the model size, precision, and batch size. Use our calculator for model-specific throughput estimates.

How much does Gaudi 2 cost per hour?

The Gaudi 2 is available starting from $1.65/hour via aws. Prices vary by provider and pricing tier (on-demand, reserved, spot). Compare pricing across all providers in the table above.