NVIDIA GPU Compute Capability Levels
This is a list of NVIDIA GPU Compute Capability levels.
The various acceleration features supported by inference engines depend on the Compute Capability level, since the hardware accelerators and dedicated functions available differ by level.
| Data Center / Professional GPUs | GeForce GPU | Capability Level | Architecture |
|---|---|---|---|
| - | GeForce RTX 5090 | 120 | Blackwell |
| - | GeForce RTX 5080 | 120 | Blackwell |
| - | GeForce RTX 5070 | 120 | Blackwell |
| - | GeForce RTX 5060 | 120 | Blackwell |
| NVIDIA B200 | - | 100 | Blackwell |
| NVIDIA B100 | - | 100 | Blackwell |
| NVIDIA H200 | - | 90 | Hopper |
| NVIDIA H100 | - | 90 | Hopper |
| NVIDIA L4 | - | 89 | Ada Lovelace |
| NVIDIA L40 | - | 89 | Ada Lovelace |
| RTX 6000 Ada Generation | - | 89 | Ada Lovelace |
| RTX 5000 Ada Generation | - | 89 | Ada Lovelace |
| RTX 4000 Ada Generation | - | 89 | Ada Lovelace |
| RTX 3000 Ada Generation | - | 89 | Ada Lovelace |
| - | GeForce RTX 4090 | 89 | Ada Lovelace |
| - | GeForce RTX 4080 | 89 | Ada Lovelace |
| - | GeForce RTX 4070 Ti / 4070 | 89 | Ada Lovelace |
| - | GeForce RTX 4060 Ti / 4060 | 89 | Ada Lovelace |
| - | GeForce RTX 4050 | 89 | Ada Lovelace |
| NVIDIA A40 | - | 86 | Ampere |
| NVIDIA A10 | - | 86 | Ampere |
| NVIDIA A16 | - | 86 | Ampere |
| NVIDIA A2 | - | 86 | Ampere |
| RTX A6000 | - | 86 | Ampere |
| RTX A5000 | - | 86 | Ampere |
| RTX A4000 | - | 86 | Ampere |
| RTX A3000 | - | 86 | Ampere |
| RTX A2000 | - | 86 | Ampere |
| RTX A1000 | - | 86 | Ampere |
| - | GeForce RTX 3090 Ti / 3090 | 86 | Ampere |
| - | GeForce RTX 3080 Ti / 3080 | 86 | Ampere |
| - | GeForce RTX 3070 Ti / 3070 | 86 | Ampere |
| - | GeForce RTX 3060 Ti / 3060 | 86 | Ampere |
| - | GeForce RTX 3050 Ti / 3050 | 86 | Ampere |
| NVIDIA A100 | - | 80 | Ampere |
| NVIDIA A30 | - | 80 | Ampere |
| NVIDIA T4 | - | 75 | Turing |
| T400 | - | 75 | Turing |
| Quadro RTX 8000 | - | 75 | Turing |
| Quadro RTX 6000 | - | 75 | Turing |
| Quadro RTX 5000 | - | 75 | Turing |
| Quadro RTX 4000 | - | 75 | Turing |
| RTX 5000 | - | 75 | Turing |
| RTX 4000 | - | 75 | Turing |
| RTX 3000 | - | 75 | Turing |
| T2000 | - | 75 | Turing |
| T1200 | - | 75 | Turing |
| T1000 | - | 75 | Turing |
| T600 | - | 75 | Turing |
| T500 | - | 75 | Turing |
| NVIDIA TITAN RTX | - | 75 | Turing |
| - | GeForce RTX 2080 Ti / 2080 Super / 2080 | 75 | Turing |
| - | GeForce RTX 2070 Super / 2070 | 75 | Turing |
| - | GeForce RTX 2060 Super / 2060 | 75 | Turing |
| - | GeForce GTX 1660 Ti / 1660 Super / 1660 | 75 | Turing |
| - | GeForce GTX 1650 Super / 1650 Ti / 1650 | 75 | Turing |
| NVIDIA V100 | - | 70 | Volta |
| Quadro GV100 | - | 70 | Volta |
| NVIDIA TITAN V | - | 70 | Volta |
