Can I run Qwen3 14B on the GeForce RTX 5090?
Runs greatQ4_K_M fits in VRAM with 20.5 GB to spare: Runs great, est. 121.3 tok/s (106.7–135.8, calibrated estimate ±12%).
32 GB VRAM · 1,792.0 GB/s · FP16 104.8 TFLOPS. Assumes 32 GB of DDR5-5600 system RAM, Windows with this GPU driving the display, 8k context and an f16 KV cache.
Every quant of Qwen3 14B on the GeForce RTX 5090
| Quant | File size | Verdict | Speed | Memory | Runs as | Notes |
|---|---|---|---|---|---|---|
| Q2_K | 5.75 GB | Runs great | est. 176.9 tok/s155.6–198.1calibrated estimate ±12% | 8.3 / 32.0 GB | Full GPU | — |
| IQ4_XS | 8.14 GB | Runs great | est. 132.3 tok/s116.4–148.2calibrated estimate ±12% | 10.7 / 32.0 GB | Full GPU | — |
| Q4_K_Mbaseline | 9.00 GB | Runs great | est. 121.3 tok/s106.7–135.8calibrated estimate ±12% | 11.5 / 32.0 GB | Full GPU | — |
| Q8_0 | 15.70 GB | Runs great | est. 73.6 tok/s64.8–82.4calibrated estimate ±12% | 18.2 / 32.0 GB | Full GPU | — |
Where the memory goes at Q4_K_M
- Weights
- 9.0 GB
- KV cache
- 1.3 GB
- Compute buffer
- 0.6 GB
- OS reserve
- 0.6 GB
- Free
- 20.5 GB
Context length vs. KV cache
| Context | KV f16 | KV q8_0 | KV q4_0 |
|---|---|---|---|
| 4k | 10.2 GBRuns great est. 129.7 tok/s114.1–145.3calibrated estimate ±12% | 9.9 GBRuns great est. 134.0 tok/s117.9–150.1calibrated estimate ±12% | 9.7 GBRuns great est. 136.5 tok/s120.1–152.9calibrated estimate ±12% |
| 8k | 10.9 GBRuns great est. 121.3 tok/s106.7–135.8calibrated estimate ±12% | 10.3 GBRuns great est. 129.1 tok/s113.6–144.6calibrated estimate ±12% | 10.0 GBRuns great est. 133.7 tok/s117.7–149.7calibrated estimate ±12% |
| 16k | 12.3 GBRuns great est. 107.4 tok/s94.5–120.2calibrated estimate ±12% | 11.1 GBRuns great est. 120.2 tok/s105.8–134.6calibrated estimate ±12% | 10.4 GBRuns great est. 128.5 tok/s113.0–143.9calibrated estimate ±12% |
| 32k | 15.2 GBRuns great est. 87.3 tok/s76.8–97.8calibrated estimate ±12% | 12.7 GBRuns great est. 105.7 tok/s93.0–118.3calibrated estimate ±12% | 11.3 GBRuns great est. 119.1 tok/s104.8–133.4calibrated estimate ±12% |
S = Runs great · A = Runs well · B = Runs slowly · F = Won't run
At Q4_K_M the verdict stays Runs great up to 32k context with an f16 KV cache, and up to 32k with q8_0.
Estimated speed
- Generation
- est. 121.3 tok/s (106.7–135.8, calibrated estimate ±12%)
- Prompt processing
- about 14,672 tok/s (8,803–20,541, rough estimate ±40%)
Measured on this exact combination
No public measurement for this exact combination yet; the numbers above are estimates.
If this is not enough
A larger model that also runs well on the GeForce RTX 5090
Runs greatQwen3.5 35B-A3B: Runs great, est. 235.0 tok/s (188.0–282.0, calibrated estimate ±20%)
A larger quant that still runs great
Runs greatQ8_0 (15.70 GB): Runs great, est. 73.6 tok/s (64.8–82.4, calibrated estimate ±12%)
How to run it
Commands for the Q4_K_M file. Both tools download from Hugging Face on first run.
File: Qwen3-14B-Q4_K_M.gguf (9.00 GB) from unsloth/Qwen3-14B-GGUF.
llama-server -hf unsloth/Qwen3-14B-GGUF:Q4_K_M -c 8192 -ngl all -fa on- -ngl all loads every layer on the GPU.
- -fa on enables flash attention (needed for KV cache quantization).
- --cache-type-k q8_0 --cache-type-v q8_0 shrinks the KV cache to 0.72 GB at 8k context.
PowerShell (quit the Ollama tray app first):
$env:OLLAMA_CONTEXT_LENGTH="8192"; ollama serve
ollama run hf.co/unsloth/Qwen3-14B-GGUF:Q4_K_M- Ollama defaults to 4k context on GPUs under 24 GB. Set OLLAMA_CONTEXT_LENGTH when starting the server, or /set parameter num_ctx inside the chat.
/set parameter num_ctx 8192
Related pages
Guide: How much VRAM do local LLMs need? 8–32 GB tiers (2026)
Other models on the GeForce RTX 5090
Qwen3 14B on other hardware
Frequently asked questions
Can I run Qwen3 14B on the GeForce RTX 5090?
Q4_K_M fits in VRAM with 20.5 GB to spare: Runs great, est. 121.3 tok/s (106.7–135.8, calibrated estimate ±12%). Q4_K_M needs 10.9 GB at 8k context; this setup has 31.4 GB of usable memory and 28.0 GB of free system RAM.
How much context can Qwen3 14B use on the GeForce RTX 5090?
At Q4_K_M the verdict stays Runs great up to 32k context with an f16 KV cache (5.4 GB of KV) and up to 32k with a q8_0 KV cache (2.9 GB). The model supports up to 32k.
Which quant should I use, and how fast is it?
Q4_K_M (9.00 GB) is the recommended balance: Runs great, est. 121.3 tok/s (106.7–135.8, calibrated estimate ±12%). The largest tracked file that still gets Runs great is Q8_0 (15.70 GB), est. 73.6 tok/s (64.8–82.4, calibrated estimate ±12%).
Speeds are estimates from memory bandwidth, calibrated against public benchmarks, and each one comes with an error band and a confidence label. Real results vary with drivers, backend, context length and thermals.