Qwen on NVIDIA Ada
Can I run Qwen3 14B on an RTX 4070?
Not at Q4_K_M — it needs 10.5 GB against 10.5 GB available. Drop to Q4_K_S and it fits, at about 34.6 tokens per second.
Won't fit
10.5 GB of 10.5 GB · 101%011 GB
Weights
8.5 GB
KV cache
1.3 GB
Runtime overhead
0.8 GB
Over the limit
0.1 GB
Short by 0.1 GB. You can run it with 39 of 40 layers on the RTX 4070 and the rest in system RAM, at roughly 28.2 tok/s — usable for batch work, painful for chat. A smaller quantisation or a shorter context is usually the better trade.
Generation28.2tok/s
Prompt processing825tok/s
Max context8Ktokens
KV per 1K tokens0GB
Every quantisation of Qwen3 14B on a RTX 4070
Highlighted row is the highest quality that still fits at 8K context.
| Quantisation | bpw | Weights | Total | Verdict | Max ctx | tok/s |
|---|---|---|---|---|---|---|
| FP16 / BF16 | 16.00 | 27.5 GB | 29.6 GB | Won't fit | — | 1.84 |
| INT8 / W8A8 | 8.50 | 14.3 GB | 16.4 GB | Won't fit | — | 5.16 |
| Q8_0 (GGUF) | 8.50 | 14.3 GB | 16.4 GB | Won't fit | — | 5.16 |
| FP8 (E4M3) | 8.00 | 13.8 GB | 15.8 GB | Won't fit | — | 5.58 |
| Q6_K | 6.56 | 11.3 GB | 13.4 GB | Won't fit | — | 8.75 |
| Q5_K_M | 5.67 | 9.7 GB | 11.8 GB | Won't fit | — | 14.3 |
| Q5_K_S | 5.52 | 9.5 GB | 11.6 GB | Won't fit | 896 | 16.0 |
| AWQ 4-bit | 4.25 | 9.4 GB | 11.5 GB | Won't fit | 1K | 16.1 |
| GPTQ 4-bit | 4.25 | 9.4 GB | 11.5 GB | Won't fit | 1K | 16.1 |
| MXFP4 | 4.25 | 9.4 GB | 11.5 GB | Won't fit | 1K | 16.1 |
| Q4_K_M | 4.85 | 8.5 GB | 10.5 GB | Won't fit | 8K | 28.2 |
| Q4_K_S | 4.58 | 8.0 GB | 10.1 GB | Fits, but tight | 10K | 34.6 |
| Q4_0 | 4.55 | 8.0 GB | 10.1 GB | Fits, but tight | 11K | 34.8 |
| IQ4_XS | 4.25 | 7.5 GB | 9.6 GB | Fits, but tight | 13K | 36.7 |
| Q3_K_M | 3.91 | 7.0 GB | 9.1 GB | Runs comfortably | 17K | 39.2 |
| IQ3_M | 3.70 | 6.7 GB | 8.8 GB | Runs comfortably | 19K | 40.9 |
| IQ3_XXS | 3.06 | 5.7 GB | 7.8 GB | Runs comfortably | 25K | 47.2 |
| Q2_K | 2.63 | 5.0 GB | 7.1 GB | Runs comfortably | 29K | 52.7 |
| IQ2_XXS | 2.06 | 4.2 GB | 6.3 GB | Runs comfortably | 35K | 62.2 |
| IQ1_M | 1.75 | 3.7 GB | 5.8 GB | Runs comfortably | 38K | 68.9 |