← Back to comparison
NVIDIA
NVIDIA GeForce RTX 4080 Super
Ada Lovelace · RTX 40 Series · ACTIVE
We earn commissions from purchases made through links on this site. This doesn't affect our rankings or recommendations.
Prices updated 5 days ago
$/GB VRAM (best)
$49/GB
MSRP: $999 (98% of MSRP)
Specifications
| VRAM | 16 GB |
| Memory type | GDDR6X |
| Bus width | 256-bit |
| Memory bandwidth | 736 GB/s |
| CUDA cores | 10,240 |
| Tensor cores | 320 |
| FP16 | 104.4 TFLOPS |
| TDP | 320W |
| Power connector | 16-pin (12VHPWR) |
| Card length | 304 mm |
| Slot width | 2.5 slots |
| PCIe | Gen 4 x16 |
| CUDA compute | 8.9 |
| Max model (Q4) | ~28B parameters |
Inference Benchmarks (Q4_K_M)
Llama 3.3 8B
95.0 tok/s
Qwen 3 32B
28.0 tok/s*
Llama 3.3 70B
—
llama.cpp, batch_size=1, ctx=4096, single GPU. Values marked with * are estimated.
What Can You Run?
| Model | Q4_K_M | Q8_0 | FP16 |
|---|---|---|---|
| Llama 3.3 8B8B | Excellent~95 tok/s | Usable | Won't fit |
| Llama 3.3 70B70.6B | Won't fit | Won't fit | Won't fit |
| Qwen 3 8B8.2B | Usable | Usable | Won't fit |
| Qwen 3 32B32.8B | Won't fit~28 tok/s | Won't fit | Won't fit |
| DeepSeek R1 70B70.6B | Won't fit | Won't fit | Won't fit |
| Mistral Nemo 12B12.2B | Usable | Usable | Won't fit |
| Phi-4 14B14B | Usable | Won't fit | Won't fit |
| Gemma 3 27B27.4B | Won't fit | Won't fit | Won't fit |
| Codestral 25B25.3B | Won't fit | Won't fit | Won't fit |
| Command R 35B35B | Won't fit | Won't fit | Won't fit |