NVIDIA B300 SXM 262GB vs NVIDIA H200 SXM 141GB

The B300 delivers 2.3x the FP8 throughput of the H200 (4,500 vs 1,979 TFLOPS dense). The B300 also carries 122 GB more memory (263 GB vs 141 GB).

B300: Blackwell Ultra, 2025 H200: Hopper, 2024
FP8 dense lead
2.3x
B300: 4,500 vs 1,979 TFLOPS
Memory
262.5 vs 141 GB
122 GB more for the B300
Bandwidth
8.0 TB/s vs 4.8 TB/s
B300 moves data faster
TDP
1.4 kW vs 700 W
H200 draws 700 W less
See rental pricing from $2.09/hr Download comparison sheet

Share this page

NVIDIA B300 SXM 262GB vs NVIDIA H200 SXM 141GB

A post with this link shows the card on the Image tab as its preview.

Printable version →

NVIDIA B300 SXM 262GB

Leads on FP8 Full specs →

What is the NVIDIA B300?

The B300 is a GPU from NVIDIA on the Blackwell Ultra architecture, an accelerator for AI training, inference and high-performance computing. It is an SXM module, the mezzanine format used in dense multi-GPU servers. B300 full specifications →

Can I rent the B300 in the cloud?

Yes. Cloud providers tracked by Flopper rent the B300 by the hour, and the pricing page compares their current rates and the price history side by side. B300 rental prices →

How does the B300 compare with other accelerators?

This page sets the B300 against the H200. The B300 is the newer generation of the two. Both are NVIDIA parts, and both ship as an SXM module, so they fit the same class of system. Flopper also compares it head to head with the B200 SXM 180GB and H100 SXM5 80GB, across every published precision, memory and power. B300 vs B200 SXM 180GB →

Where do the B300 figures come from?

Every B300 figure on this page traces to NVIDIA's own datasheet or specification table. How Flopper verifies specifications →

NVIDIA H200 SXM 141GB

Full specs →

What is the NVIDIA H200?

The H200 is a GPU from NVIDIA on the Hopper architecture, an accelerator for AI training, inference and high-performance computing. It is an SXM module, the mezzanine format used in dense multi-GPU servers. H200 full specifications →

Can I rent the H200 in the cloud?

Yes. Cloud providers tracked by Flopper rent the H200 by the hour, and the pricing page compares their current rates and the price history side by side. H200 rental prices →

How does the H200 compare with other accelerators?

This page sets the H200 against the B300. The H200 is the older generation of the two. Both are NVIDIA parts, and both ship as an SXM module, so they fit the same class of system. Flopper also compares it head to head with the H100 SXM5 80GB, A100 SXM4 80GB and Instinct MI300X 192GB, across every published precision, memory and power. H200 vs H100 SXM5 80GB →

Where do the H200 figures come from?

Every H200 figure on this page traces to NVIDIA's own datasheet or specification table. How Flopper verifies specifications →

Specifications

B300
H200
Architecture
Blackwell UltraHopper
Launch Year
20252024
Form Factor
SXMSXM
Memory
263 GB141 GB
Memory Bandwidth
8.0 TB/s4.8 TB/s
TDP
1.4 kW700 W
Process Node
4nm4nm

Performance (TFLOPS)

B300
H200
1.3 TFLOPS 34 TFLOPS
75 TFLOPS 67 TFLOPS
1,125 TFLOPS
2,250 TFLOPS sparse
494 TFLOPS
989 TFLOPS sparse
2,250 TFLOPS
4,500 TFLOPS sparse
990 TFLOPS
1,979 TFLOPS sparse
2,250 TFLOPS
4,500 TFLOPS sparse
990 TFLOPS
1,979 TFLOPS sparse
4,500 TFLOPS
9,000 TFLOPS sparse
1,979 TFLOPS
3,958 TFLOPS sparse
FP6
4,500 TFLOPS
9,000 TFLOPS sparse
No verified data available
13,500 TFLOPS
18,000 TFLOPS sparse
No verified data available
188 TOPS
375 TOPS sparse
1,979 TOPS
3,958 TOPS sparse

Figures are dense. Where a vendor publishes a figure it sits under the dense one and never decides a winner. Hover any precision for its definition, or read the GPU glossary.

FLOPS by Precision

What actually differs

The B300 is the newer part: Blackwell Ultra (2025) against the H200's Hopper (2024).

The B300 carries 122 GB more memory (263 GB against 141 GB), so larger models fit on one GPU before you shard. Bandwidth is 8.0 TB/s against 4.8 TB/s, which sets the ceiling on decode speed.

Power is 1.4 kW against 700 W, and at FP32 the H200 does more work per watt. Weight the FP8 and FP16 rows for LLM work and the FP64 row for HPC.

New to the terms? The glossary explains FP64, FP32, TF32, BF16, FP16, FP8, FP4, INT8, Sparsity and TFLOPS. For how the generations relate, see NVIDIA architectures explained.

Frequently asked questions

What is better, the B300 or the H200?

The B300, on throughput. At FP8 precision the B300 reaches 4,500 TFLOPS dense against 1,979 TFLOPS for the H200. The performance table on this page lists every published precision for both GPUs.

Which should you choose, the B300 or the H200?

Choose the B300 for training and high-throughput inference: it has 2.3x the FP8 throughput of the H200. It also carries 122 GB more memory, so larger models fit on one GPU before you shard. The B300 draws 700 W more power, so budget for cooling and power delivery before you commit.

Which has more memory, the B300 or the H200?

The B300 carries 263 GB of memory versus 141 GB for the H200. Memory bandwidth is 8.0 TB/s for the B300 and 4.8 TB/s for the H200.

How much power do the B300 and the H200 draw?

The B300 is rated at 1.4 kW TDP and the H200 at 700 W. On FP32 throughput per watt, the H200 is the more efficient part.

How much do the B300 and the H200 cost to rent?

Live cloud listings tracked by Flopper price the B300 from $4.08 per GPU-hour across 19 providers, and the H200 from $2.09 per GPU-hour across 31 providers. The rental section on this page links to every provider, rate and the price history for each GPU.

Where to Rent

NVIDIA B300 SXM 262GB

Rent from

$4.08 / GPU-hr

Lowest of 24 live listings across 19 providers · median $7.92

See all B300 rental prices →

Price history and every provider on the pricing page. Some links are affiliate links.

NVIDIA H200 SXM 141GB

Rent from

$2.09 / GPU-hr

Lowest of 54 live listings across 31 providers · median $4.47

See all H200 rental prices →

Price history and every provider on the pricing page. Some links are affiliate links.

Get Comparison Updates

New GPUs added weekly. Be the first to see how they compare.

© 2026 Flopper.io - Compare the GPUs Powering AI