RunsOnMy ModelsGPUs & MacsBy VRAMBest GPUCan I run…?OllamaStart hereAPI pricesVRAM calculatorNew

Can DeepSeek R1 0528 Qwen3 8B run on the RTX 5090?

Yes. At 8K context the best fit is F16, using about 17.1 GB of the 30.4 GB usable on the RTX 5090. Estimated speed ceiling: ~109 tokens/sec.

Memory by quant and context

Quant4K ctx8K ctx32K ctx128K ctxMax tok/s (est., 8K)
2-bit (Q2_K) •4.3 GB Fits4.9 GB Fits8.2 GB Fits21.7 GB Fits546
3-bit (Q3_K_M) •5.1 GB Fits5.7 GB Fits9.0 GB Fits22.5 GB Fits435
4-bit (Q4_K_M) •5.9 GB Fits6.5 GB Fits9.9 GB Fits23.4 GB Fits356
5-bit (Q5_K_M) •6.7 GB Fits7.3 GB Fits10.6 GB Fits24.1 GB Fits306
6-bit (Q6_K) •7.5 GB Fits8.1 GB Fits11.5 GB Fits25.0 GB Fits266
8-bit (Q8_0) •9.4 GB Fits9.9 GB Fits13.3 GB Fits26.8 GB Fits206
16-bit (F16/BF16) •16.5 GB Fits17.1 GB Fits20.5 GB Fits34.0 GB No109

• real GGUF file size from bartowski/deepseek-ai_DeepSeek-R1-0528-Qwen3-8B-GGUF. Speed is a bandwidth ceiling (1792 GB/s), not a benchmark. Methodology.

What fits instead

"Runs" and "fits" mean the whole model in GPU memory at 8K context (4-bit or better for the model pick). Prices are US launch MSRPs, not today's street prices. "Popular" means the 60 most-downloaded models we track.

More

DeepSeek R1 0528 Qwen3 8B on every GPU and Mac · Every model that runs on the RTX 5090 · Other models that fit 32 GB VRAM · Try other settings in the calculator

Guides: Best GPU for local LLMs by budget · Can I run gpt-oss, Llama 70B or Qwen3? · Mac vs PC for local LLMs

Data: Hugging Face Hub (deepseek-ai/DeepSeek-R1-0528-Qwen3-8B), NVIDIA specs. Model data last checked 2026-11-18; refreshed hourly.