Side-by-side GPU comparison: specs, memory, compute performance, and live cloud pricing.
Verdict
RTX 5090 has 16 GB more VRAM, making it better suited for large models and long context windows. For compute-bound workloads like training, RTX 5090 delivers 12.0× higher FP16 throughput. RTX 5090 supports a broader range of models (577 vs 502 from this catalog), giving more flexibility.
Specifications
| A16 | RTX 5090 | |
|---|---|---|
| VRAM | 16 GB | 32 GB |
| VRAM Type | GDDR6 | GDDR7 |
| Memory Bandwidth | 0.4 TB/s | 1.8 TB/s |
| FP16 Performance | 18 TFLOPS | 210 TFLOPS |
| Manufacturer | NVIDIA | NVIDIA |
| FP8 Support | No | Yes |
| FP4 Support | No | Yes |
Price / Performance
Based on cheapest single-GPU on-demand pricing. Lower $/TFLOP = better compute value; lower $/GB = better memory value.
| A16 | RTX 5090 | |
|---|---|---|
| $/hr (cheapest) | — | $0.44 |
| $/TFLOP (compute value) | — | $0.0021 |
| $/GB VRAM (memory value) | — | $0.0138 |
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.
A16 (502 models)
RTX 5090 (577 models)
You might also compare…
Pricing data refreshed hourly · Last updated April 11, 2026 · Browse all comparisons