What's the price difference between NVIDIA B300 and NVIDIA H100?

On Runcrate today, NVIDIA B300 starts at $16.77/hr and NVIDIA H100 starts at $2.49/hr; NVIDIA H100 is cheaper by 85%.

Can I deploy NVIDIA B300 or NVIDIA H100 on Runcrate today?

NVIDIA B300 is on demand in 1 region from $16.77/hr, billed by the second on most GPUs. NVIDIA H100 is on demand in 6 regions from $2.49/hr, billed by the second on most GPUs.

How much VRAM does each GPU have?

NVIDIA B300 has 288GB HBM3e. NVIDIA H100 has 80GB HBM3.

B300 or H100: which holds a 70B model in 8-bit weights?

It needs 70 GB of 8-bit weights. On the B300: 1 GPU, 218 GB spare. On the H100: 1 GPU, 10 GB spare. Pick the B300: same GPU count, up to 2.4x the memory bandwidth.

GPU comparison

B300 vs H100.

The B300 is ahead on memory (288 GB vs 80 GB) and speed (at least 2.2x the FP16 throughput and up to 2.4x the memory bandwidth). The H100 makes sense when your model fits in 80 GB and its hourly price is lower.

Launch H100 Reserve a cluster

Rent either on Runcrate

B300from $16.77/hr
$0.058 per GB/hr$0.75 per 100 TFLOPS/hr
Launch B300
H100Cheaperfrom $2.49/hr
$0.031 per GB/hrper TFLOPS: varies by card
Launch H100

Live hourly price, then per GB of memory and per 100 FP16 TFLOPS.

Which is better for

Which to pick for each job. Memory fit and throughput, from the datasheets.

JobB300H100Pick

Hold a 70B model in 8-bit weights
70 GB of 8-bit weights
B300: 1 GPU, 218 GB spare
H100: 1 GPU, 10 GB spare
B300. Same GPU count, up to 2.4x the memory bandwidth.
Serve an 8B model in 16-bit weights
16 GB of 16-bit weights
B300: 1 GPU, 272 GB spare
H100: 1 GPU, 64 GB spare
B300. Same GPU count, up to 2.4x the memory bandwidth.
Fine-tune a 70B model with QLoRA
About 46 GB with 4-bit weights
B300: 1 GPU, 242 GB spare
H100: 1 GPU, 34 GB spare
B300. Same GPU count, at least 2.2x the FP16 throughput.
Train on a full node
FP16 tensor throughput per node
B300: 8x, 18,000 TFLOPS, NVLink
H100: 8x, up to 7,912 TFLOPS
B300. at least 2.2x the FP16 throughput per GPU.
Throughput per watt
FP16 TFLOPS per 100 W of TDP
B300: Varies by card
H100: Varies by card
Either. Varies by card.

Compare another pair

Blackwell

NVIDIA B300

from$16.77/hr

Available now · 1 region

Memory: 288 GB HBM3e
FP16 tensor: 2250 TFLOPS
Bandwidth: up to 8.0 TB/s

Launch B300 Reserve a cluster

Full pricing details

Hopper

NVIDIA H100

from$2.49/hrCheaper

Available now · 6 regions

Memory: 80 GB HBM3
FP16 tensor: up to 989 TFLOPS
Bandwidth: up to 3.4 TB/s

Launch H100 Reserve a cluster

Full pricing details

At a glance

Spec by spec. The longer bar is the better number.

Memory

NVIDIA B300288 GB

NVIDIA H10080 GB

Memory bandwidth

NVIDIA B300up to 8.0 TB/s

NVIDIA H100up to 3.4 TB/s

FP16 tensor

NVIDIA B3002250 TFLOPS

NVIDIA H100up to 989 TFLOPS

Hourly price, from (live)

NVIDIA B300$16.77/hr

NVIDIA H100$2.49/hr

Full specs

The complete spec sheet.

Spec	NVIDIA B300	NVIDIA H100
Architecture	Blackwell	Hopper
Memory	288 GB	80 GB
Memory type	HBM3e	HBM3
Memory bandwidth	up to 8.0 TB/s	up to 3.4 TB/s
FP32	75 TFLOPS	up to 67 TFLOPS
FP16 tensor	2250 TFLOPS	up to 989 TFLOPS
INT8	187.5 TOPS	up to 2400 TOPS
TDP	up to 1400 W	up to 700 W
GPUs per node	8	8
Hourly price (live)	from $16.77/hr	from $2.49/hr

Questions

B300 or H100. What buyers ask first.

Other comparisons