Can an RTX 4060 run Qwen3 32B?
Partly
Not entirely. About 19 of its 64 layers fit on the card and the rest run on the CPU, which brings generation down to roughly 2.1 tokens per second.
The numbers
Qwen3 32B has 32.76 billion parameters across 64 layers. The GeForce RTX 4060 has 8 GB of VRAM at 272 GB/s, of which about 7.2 GB is left for a model once the operating system has taken its share.
| Quantisation | Total needed | Fits | Speed |
|---|---|---|---|
| Q8_0 | 35.33 GB | no | — |
| Q6_K | 27.93 GB | no | — |
| Q5_K_M | 24.61 GB | no | — |
| Q4_K_M | 21.33 GB | no | — |
| IQ4_XS | 19.12 GB | no | — |
| Q3_K_M | 17.83 GB | no | — |
| IQ3_XXS degraded | 14.58 GB | no | — |
| Q2_K degraded | 15.69 GB | no | — |
| IQ2_XXS degraded | 10.77 GB | no | — |
Nothing on this ladder fits, including the two-bit levels we do not recommend.
What to do instead
The straightforward answer is a smaller model from the same family. These do fit on a GeForce RTX 4060, at a quantisation worth using:
- Qwen3.5 9B — 9.65B at IQ4_XS, 6.63 GB, about 38.6 tokens/s
- Qwen2.5 VL 7B Instruct — 8.29B at Q5_K_M, 6.75 GB, about 37.6 tokens/s
- Qwen3 8B — 8.19B at Q4_K_M, 6.58 GB, about 38.9 tokens/s
You can also run it as is. Ollama and llama.cpp will put 19 of the 64 layers on the card and the rest on the CPU without being asked. It works, at roughly 2.1 tokens per second, which is fine for a batch job and tiring for a conversation.
Shortening the context window will not rescue this one: at 4k it still wants 20.18 GB against 21.33 GB at 8k, because the weights rather than the cache are what fill the card. The gap here is too large for settings to close.
Or the hardware that does run it
The cheapest device we track that runs Qwen3 32B at a quantisation worth using is the Radeon RX 7900 XT, around $650 , running it at IQ4_XS at about 31.6 tokens per second. Check the current price (affiliate link; indicative price reviewed 2026-09-01). Compare every option.
See how fast it feels
GeForce RTX 4060 running Qwen3 32B, 19 of 64 layers on the GPU
YouWhy does my model use more memory when the conversation gets longer?
Because of the KV cache. Every token you send leaves behind a key and a value vector in each layer of the model, and those stay in memory for as long as the conversation lasts.
The weights are a fixed cost: load a 4-bit 8B model and that is about 4.8 GB, whether you write one word or ten thousand. The cache is the part that grows, and it grows in a straight line with the number of tokens in the window.
How steeply depends on the model's attention design. With grouped-query attention, several query heads share one key-value pair, which cuts the cache by that ratio. Without it, every head keeps its own, and a long context can cost more memory than the weights themselves.
If you are short on memory, the first thing to try is lowering the context window in your runtime, not the quantisation.
Reading your question
Simulated from our estimate at a 512-token question, not a recording. Assumes nothing else is competing for the GPU. The same answer on RTX 4090 at Q4_K_M would take about 4.5 s.
Keep going
Everything the GeForce RTX 4060 runs, the full breakdown for Qwen3 32B, or check a different pairing in the calculator.