Side-by-side GPU comparison: specs, memory, compute performance, and live cloud pricing.
Verdict
H100 SXM has 32 GB more VRAM, making it better suited for large models and long context windows. For compute-bound workloads like training, H100 SXM delivers 2.7× higher FP16 throughput. At $0.60/hr vs $1.80/hr, L40S is the more cost-efficient choice for inference. H100 SXM supports a broader range of models (636 vs 624 from this catalog), giving more flexibility.
Specifications
| H100 SXM | L40S | |
|---|---|---|
| VRAM | 80 GB | 48 GB |
| VRAM Type | HBM3 | GDDR6 |
| Memory Bandwidth | 3.4 TB/s | 0.9 TB/s |
| FP16 Performance | 990 TFLOPS | 366 TFLOPS |
| Manufacturer | NVIDIA | NVIDIA |
| FP8 Support | Yes | Yes |
| FP4 Support | No | No |
Price / Performance
Based on cheapest single-GPU on-demand pricing. Lower $/TFLOP = better compute value; lower $/GB = better memory value.
| H100 SXM | L40S | |
|---|---|---|
| $/hr (cheapest) | $1.80 | $0.60✓ best |
| $/TFLOP (compute value) | $0.0018 | $0.0016✓ best |
| $/GB VRAM (memory value) | $0.0225 | $0.0125✓ best |
Cloud Pricing
Cheapest on-demand price per provider (single GPU).
Model Compatibility
Models from the catalog that fit on each GPU, grouped by required precision.
H100 SXM (636 models)
L40S (624 models)
You might also compare…
Pricing data refreshed hourly · Last updated April 11, 2026 · Browse all comparisons