self-hosted/ai
§01·spec · /gpus

RTX 4070 Ti Super

nvidia40 series16GB VRAM
§02·models that run on this GPU
3 total
ModelVerticalBest speedMin VRAMWorksBenchmarks
gpt-oss 20Bllm128.9tokens/s16GB2check ↗
Qwen3-8Bllm96.3tokens/s16GB2check ↗
Qwen3 14Bllm58.1tokens/s16GB2check ↗