Best local AI models for AMD R9 390X

The largest models that fit fully

The 30 largest of the 123 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 Radeon R9 390X graphics card features 8 GB of GDDR5 video memory. This onboard memory determines the size of the artificial intelligence models you can run entirely on the hardware. When a model fits completely within this 8 GB limit, the graphics processor handles all calculations. This local execution ensures the fastest possible generation speeds for text and images.

To fit larger models into the 8 GB limit, developers use quantization. The quantization column shows the compression level applied to each model. For example, Gemma 2 9B fits at the Q4_K_M quantization level which uses exactly 8 GB of video memory. Models like Llama 3.1 8B can run at a higher quality Q5_K_M quantization level because they only require 7.4 GB of video memory. Smaller models like Qwen2.5 7B can run at the even higher Q6_K quantization level while using only 6.9 GB of video memory.

If a model exceeds the 8 GB video memory limit, you must use CPU offload. This process splits the model between your graphics card and your system memory. We assume your computer has 32 GB of system RAM for these scenarios. For instance, running Mistral NeMo 12B at the Q4_K_M quantization level requires 8.8 GB of video memory and 10.8 GB of system RAM. Running FLUX.1 dev at FP8 requires 14.4 GB of video memory and 16.4 GB of system RAM.

CPU offload comes with a significant performance cost. Sending data between the graphics card and system RAM over the system bus is much slower than using the GDDR5 memory directly. While offloading allows you to run larger models like Vicuna 13B or Gemma 3 12B, your generation speeds will drop. For the best performance, choose models that fit entirely within the 8 GB video memory of your card.

You must also consider the memory cost of context. 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 model will require more memory. This extra memory usage might force a model that normally fits in your 8 GB video memory to offload parts of its workload to your system RAM.