Legacy Budget Desktop (GTX 1650 4GB)

Scored against every model in our directory at Q4_K_M, using the ~4GB of VRAM on this machine and its 128 GB/s of memory bandwidth. Of 30 models, 1 runs comfortably, 0 are a tight fit, 17 need CPU offload, and 12 will not load. The largest model that fits well here is Meta Llama 3.2 3B Instruct.

How these are estimated: tokens/sec ≈ memory bandwidth × 0.65 ÷ active weights in GB. These are arithmetic estimates from published specifications, not lab measurements — BFWR does not run a hardware test lab. Real throughput varies with runner, quantization, context length and thermals.

Diffusion models size up differently from language models: the text encoder is often as large as the image model itself, and memory scales with resolution. Measured at 1024×1024, one image at a time, against the ~4GB of VRAM here, 2 of 11 open image models will run — the best being Stable Diffusion XL 1.0 at NF4 / INT4 with text encoder on cpu, peaking near 4.0GB.

Weight sizes are read from the real .safetensors files in each model’s Hugging Face repository. Activation memory is an approximation that varies with the attention implementation and sampler, so leave headroom. These are planning figures, not benchmarks.

What is the largest model the Legacy Budget Desktop (GTX 1650 4GB) can run?

Meta Llama 3.2 3B Instruct (3.21B) is the largest model that fits, at roughly 2GB in Q4_K_M against 4GB of VRAM. It is rated "Comfortable" here, with an estimated 41.6 tokens/sec.

How many tokens per second should I expect?

Speed depends on the model, not just the machine. On this hardware the estimates range from about 41.6 tokens/sec for the smallest models down to single digits for the largest that still fit. The estimate uses memory bandwidth (128 GB/s here) × 0.65 ÷ active weights in GB, and is not a measured benchmark.

Can the Legacy Budget Desktop (GTX 1650 4GB) run a 70B model?

Not in a usable way. A 70B model needs roughly 40GB+ at Q4_K_M before context, which exceeds the 4GB of VRAM available here. A mixture-of-experts model is usually the better route on this machine, since only a fraction of its weights are active per token.

How much context can I use?

8k tokens is a sensible ceiling on this machine at Llama 3.2 1B / 3B only. Context is stored in the KV cache, which grows with every token and competes with the model weights for the same memory, so a longer context means a smaller model.

Can the Legacy Budget Desktop (GTX 1650 4GB) generate images with Flux or Stable Diffusion?

Yes. 2 of 11 open image models fit at 1024x1024, the best being Stable Diffusion XL 1.0 at NF4 / INT4 with text encoder on cpu, peaking near 4.0GB. Higher resolutions and larger batches cost proportionally more.

Which runner works best on the Legacy Budget Desktop (GTX 1650 4GB)?

Ollama are all expected to run on this hardware. Ollama is the simplest starting point, and `ollama run gemma4:e4b` is a good first model for this machine.

In-depth reads that pair with this machine — sizing memory, picking a runner, and choosing quantization for ~4GB of usable memory.

• Best Budget GPUs for Local LLMs16GB cards, used RTX 3090s and current street prices.Read the guide
• RTX 5080 vs RTX 4090: Why 16GB VRAM Is the LimitWhy VRAM capacity beats raw speed for local inference.Read the guide
• How Much RAM and VRAM Do You Need to Run Local LLMs?Model weights, KV cache and OS overhead, tier by tier.Read the guide