Best local AI models for NVIDIA GT 740A

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 NVIDIA GT 740A is an entry level graphics card equipped with 2 GB of DDR3 video memory. This hardware configuration places strict limits on the size of local artificial intelligence models you can run. To load a model entirely onto the graphics processor, the model files and working memory must fit within this 2 GB boundary. Running models locally on this card requires careful selection of compact architectures and optimized quantization levels.

Quantization is a compression method that reduces the precision of model weights to save memory. The quant column indicates the best balance of size and quality for this hardware. For example, the 2.8B Allegro model and the 2.7B Open-Sora Plan model can run using the Q4_K_M quantization level, which uses 2 GB of video memory. Smaller models like the 1.7B Qwen3 or the 1.7B SmolLM2 can run at a higher quality Q6_K quantization level while using 1.7 GB of memory.

When a model exceeds the 2 GB video memory limit, you must use CPU offload. This technique splits the workload between your graphics card and your system RAM. If you have 32 GB of system RAM, you can run larger models like the 3.5B SDXL Turbo or the 3.5B SDXL Lightning. These models need 2.6 GB of video memory at Q4_K_M quantization and require an additional 4.6 GB of system RAM to function.

Using CPU offload comes with a significant performance cost. Transferring data between the DDR3 video memory and the system RAM is much slower than running calculations entirely on the graphics card. While offloading allows you to run models like the 3.417B Stable Diffusion XL or the 3.3B MusicGen, the generation times will be noticeably longer. For the fastest response times, you should select models that fit completely within the 2 GB video memory limit.

You must also consider the memory cost of context length when running text models. The listed memory usage figures assume a standard starting context. As you input longer prompts or generate longer responses, the memory required by the context window grows. Running a model close to the 2 GB limit, such as the 1.1B TinyLlama using 1.4 GB at Q8_0, leaves some room for context. Running a 2B model at its limit will leave very little room for extended conversations before running out of memory.