text model · Nemotron · Windows
Can I run Llama-3.3-Nemotron-Super-49B-v1 on AMD Radeon RX 7900 XTX (24GB)?
No. Llama-3.3-Nemotron-Super-49B-v1 needs ~30.6 GB even at Q4_K_M, but AMD Radeon RX 7900 XTX (24GB) only has ~23 GB usable.
Needs ~30.6 GB even at Q4_K_M, but only ~23 GB is usable.
That figure is at a 4k context and moves about ±15% as context length changes.
The gap is about 7.6 GB: Llama-3.3-Nemotron-Super-49B-v1 needs roughly 30.6 GB at Q4_K_M and AMD Radeon RX 7900 XTX (24GB) leaves only about 23 GB usable for a model. The lightest tracked hardware that runs Llama-3.3-Nemotron-Super-49B-v1 is the Nvidia GeForce RTX 5090 (32GB) at 32 GB. See Llama-3.3-Nemotron-Super-49B-v1 on Nvidia GeForce RTX 5090 (32GB).
Too big for AMD Radeon RX 7900 XTX (24GB)'s VRAM, but you could offload some layers to system RAM and still run it at roughly ~5 tok/s, far slower than a model that fits. How this is estimated.
- Q4_K_M needed
- ~30.6 GB
- Usable on device
- ~23 GB
- Device memory
- 24 GB
Which quant fits
How to run it
On Windows use LM Studio (Best GUI on Windows, auto-detects CUDA/Vulkan backends.).
- Parameters
- 49B
- Q4_K_M size
- 28.14 GB
- Q8_0 size
- 49.36 GB
- Context
- 128k
- Memory
- 24 GB vram
- Usable for weights
- ~23 GB
- Power draw
- ~355 W
- Best runtime
- Ollama (ROCm) / llama.cpp ROCm (Linux)
What you can run instead
Run Llama-3.3-Nemotron-Super-49B-v1 on other hardware
FAQ
Can AMD Radeon RX 7900 XTX (24GB) run Llama-3.3-Nemotron-Super-49B-v1?
No. Llama-3.3-Nemotron-Super-49B-v1 needs ~30.6 GB even at Q4_K_M, but AMD Radeon RX 7900 XTX (24GB) only has ~23 GB usable.
How much memory does Llama-3.3-Nemotron-Super-49B-v1 need?
AMD Radeon RX 7900 XTX (24GB) does not have enough memory. At Q4_K_M the weights are ~28.14 GB; with KV cache and runtime overhead, budget ~30.6 GB at a 4k context.
What is the best tool to run Llama-3.3-Nemotron-Super-49B-v1 on Windows?
LM Studio for a simple setup; Ollama (CUDA) for the most speed. AMD GPUs run via Vulkan/ROCm at roughly half CUDA throughput. NVIDIA is the smooth path on Windows.
Embed this
A live badge for your model card or README, updated as the data is. [](https://localmodel.run/can-i-run/nemotron-super-49b/amd-rx-7900-xtx-24gb) Sources
Weights are measured from GGUF files; KV cache and overhead are computed, so totals can vary ~15% with context and runtime. Any tok/s is a bandwidth estimate. See methodology.