NVIDIA L20 48GB vs NVIDIA GeForce RTX 4090 24GB

The GeForce RTX 4090 delivers 38% more FP8 throughput than the L20 (330 vs 239 TFLOPS dense). The L20 counters with 24 GB more memory (48 GB vs 24 GB).

L20: Ada Lovelace, 2024 GeForce RTX 4090: Ada Lovelace, 2022
FP8 dense lead
+38%
GeForce RTX 4090: 330 vs 239 TFLOPS
VRAM
48 vs 24 GB
24 GB more for the L20
Bandwidth
864 GB/s vs 1.0 TB/s
GeForce RTX 4090 moves data faster
TDP
275 W vs 450 W
L20 draws 175 W less

Specifications

L20 GeForce RTX 4090
Architecture
Ada Lovelace Ada Lovelace
Launch Year
2024 2022
Form Factor
PCIe PCIe
VRAM
48 GB MORE24 GB
Memory Bandwidth
864 GB/s 1.0 TB/s FASTER
TDP
275 W 450 W
Process Node
4nm

Performance (TFLOPS)

L20 GeForce RTX 4090
FP64
1.3 TFlops
FP32
59.8 TFlops 82.6 TFlops BEST
FP16
119.5 TFlops
sparse not published
165.2 TFlops BEST
sparse not published
BF16
119.5 TFlops
sparse not published
165.2 TFlops BEST
sparse not published
FP8
239.0 TFlops
sparse not published
330.3 TFlops BEST
sparse not published
INT8
239.0 TFlops
sparse not published
330.3 TFlops BEST
sparse not published

FLOPS by Precision

What actually differs

The L20 is the newer part: Ada Lovelace, launched in 2024, against the GeForce RTX 4090's Ada Lovelace from 2022. Newer architectures typically add lower-precision formats and better throughput per watt, so check the precision rows your workload actually uses.

Dense throughput for the L20 against the GeForce RTX 4090: FP32 60 vs 83 TFLOPS, FP16 120 vs 165 TFLOPS, FP8 239 vs 330 TFLOPS. Sparse figures, where the vendor publishes them, appear under each dense number in the performance table.

Memory is 48 GB against 24 GB, fed at 864 GB/s versus 1.0 TB/s. More memory per GPU means larger models fit before you have to shard across cards, which often matters more than raw TFLOPS for inference.

Power budgets are 275 W for the L20 and 450 W for the GeForce RTX 4090. At FP32 that works out to 0.22 against 0.18 TFLOPS per watt.

For LLM training and inference, weight the FP8 and FP16 rows and memory capacity most heavily. For scientific and HPC workloads, the FP64 row is the one to read.

Frequently asked questions

Is the L20 faster than the GeForce RTX 4090?

At FP8 precision the GeForce RTX 4090 reaches 330 TFLOPS dense against 239 TFLOPS for the L20. The performance table on this page lists every published precision for both GPUs.

Which has more memory, the L20 or the GeForce RTX 4090?

The L20 carries 48 GB of VRAM versus 24 GB for the GeForce RTX 4090. Memory bandwidth is 864 GB/s for the L20 and 1.0 TB/s for the GeForce RTX 4090.

How much power do the L20 and the GeForce RTX 4090 draw?

The L20 is rated at 275 W TDP and the GeForce RTX 4090 at 450 W. On FP32 throughput per watt, the L20 is the more efficient part.

Can I rent the L20 or the GeForce RTX 4090 in the cloud?

Yes. Live cloud listings tracked by Flopper start at $0.13 per GPU hour. The rental pricing section on this page lists current providers and rates for both GPUs.

Where to Rent

NVIDIA GeForce RTX 4090 24GB

ProviderInstancePrice/GPU-hrChecked
Vast.ai
$0.13 5h ago View →
RunPod
RTX 4090 (Community Cloud) $0.34 5h ago View →
Excess Supply
auto-rtx4090-8x-nvidia-geforce-rtx-us-central-1-6db4066c7c $0.40 5h ago View →
RunPod
RTX 4090 (Secure Cloud) $0.74 5h ago View →
All GeForce RTX 4090 listings and price history →

Get Comparison Updates

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