Home › GPUs & Macs › NVIDIA RTX 5070 Ti vs NVIDIA RTX 5080
NVIDIA RTX 5070 Ti vs NVIDIA RTX 5080 for local LLMs
Across 5 tests measured on both, the two are about even (median ratio 1.04×).
On Llama 2 7B (Q4_0, llama-bench) the NVIDIA RTX 5070 Ti generates 182 tokens/s and the NVIDIA RTX 5080 192 tokens/s (measured).
Same model, same quant, same test
| Model | Quant / test | RTX 5070 Ti tok/s | RTX 5080 tok/s | Faster | Sources and run details |
|---|---|---|---|---|---|
| Llama 2 7B | Q4_0 llama-bench | 182 CUDA | 192 Vulkan | RTX 5080 +5% | llama.cpp #15013 RTX 5070 Ti · llama.cpp #10879 RTX 5080 |
| gpt-oss 20B | MXFP4 llama-bench | 189 CUDA | 229 CUDA | RTX 5080 +21% | llama.cpp #15396 RTX 5070 Ti · llama.cpp #15396 RTX 5080 |
| Llama 3.1 8B Instruct | Q4_K_M LocalScore | 50.1 llamafile | 35.5 llamafile | RTX 5070 Ti +41% | LocalScore RTX 5070 Ti · LocalScore RTX 5080 |
| Llama 3.2 1B Instruct | Q4_K_M LocalScore | 145 llamafile | 140 llamafile | RTX 5070 Ti +4% | LocalScore RTX 5070 Ti · LocalScore RTX 5080 |
| Qwen2.5 14B Instruct | Q4_K_M LocalScore | 31.3 llamafile | 29.3 llamafile | RTX 5070 Ti +7% | LocalScore RTX 5070 Ti · LocalScore RTX 5080 |
Generation speed in tokens/s, best measured backend for each chip (shown under the number). Backends can differ between the two chips; check them before drawing conclusions. Measured results only, nothing estimated.
Specs
| NVIDIA RTX 5070 Ti | NVIDIA RTX 5080 | |
|---|---|---|
| Memory | 16 GB | 16 GB |
| Memory bandwidth | 896 GB/s | 960 GB/s |
All results for the NVIDIA RTX 5070 Ti · All results for the NVIDIA RTX 5080 · More comparisons