NVIDIA L4
NVIDIA · Ada Lovelace
Shipping (widely available)
inferenceavailablelow-poweredge
Memory
24GB GDDR6
FP8 TFLOPS
242
Power
72W
Released
Jan 1, 2023
Technical Specifications
| Manufacturer | NVIDIA |
|---|---|
| Architecture | Ada Lovelace |
| Memory | 24GB GDDR6 |
| Memory Bandwidth | 0.300 TB/s |
| FP16 TFLOPS | 121 |
| FP8 TFLOPS | 242 |
| Interconnect | PCIe Gen4 |
| Power (TDP) | 72W |
| Released | Jan 1, 2023 |
| Status | Shipping (widely available) |
Availability
Widely available — low-power inference GPU for edge and cloud
Pricing
Retail (estimated)
2000-3500
Cloud Pricing ($/hr)
| gcp_t4_l4 | $0.75/hr |
|---|---|
| runpod | $0.49/hr |
Key Specs
Memory: 24GB GDDR6
Bandwidth: 0.300 TB/s
FP8: 242 TFLOPS
Power: 72W