Best local AI models for AMD HD 7870 XT

The largest models that fit fully

The 30 largest of the 56 models that fit; every smaller model in the catalog fits too. Best quant means the highest quality compression whose weights and 4k context both sit inside the memory.

Close, but only with CPU offload

These need more than the card holds at their smallest practical quant, so part of the model runs from system memory (figures assume 32 GB of it). They work, several times slower.

How to read this

The AMD HD 7870 XT graphics card comes equipped with 2 GB of GDDR5 video memory. This memory size is the absolute limit for running local AI models entirely on your hardware. To fit within this tight budget, models must be compressed using quantization. The quant column shows the specific compression level needed to run each model. For example, Allegro 2.8B and Open-Sora Plan 2.7B require a Q4_K_M quantization to fit exactly into 2 GB of video memory.

Smaller models can run with much less compression on this hardware. Models like SmolLM2 1.7B and Qwen3 1.7B can use a Q6_K quantization which uses 1.7 GB of video memory. Even smaller models like TinyLlama 1.1B and SantaCoder 1.1B can run at Q8_0 quantization while using only 1.4 GB of video memory. Higher quantization levels preserve more of the original model quality but require more memory space.

If you want to run larger models, you must use CPU offloading. This process splits the model between your graphics card and your system memory. Offloading allows you to run models that exceed your 2 GB limit, but it comes with a heavy speed cost. Processing data across the system memory bus is much slower than running everything directly on your graphics card video memory.

With a system containing 32 GB of system RAM, you can offload models like SmolLM3 3B or Kandinsky 3.1. These models require 2.2 GB of video memory at Q4_K_M quantization and will use an additional 4.2 GB of system RAM. You can also run larger media models like Stable Diffusion XL which requires 4.1 GB of video memory at FP8 or optimized settings along with 6.1 GB of system RAM.

When running local text models, you must also consider the context window. The memory figures listed here assume a standard 4k context window. If you increase the context window to process longer documents or chat histories, the memory usage will rise quickly. This extra memory demand can easily push a model past your 2 GB limit and force your system to slow down.