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

Can Granite 3.3 2B Instruct run on the Ryzen AI Max+ 395 128GB?

Yes. At 8K context the best fit is F16, using about 5.8 GB of the 109 GB usable on the Ryzen AI Max+ 395 128GB. Estimated speed ceiling: ~51 tokens/sec.

Memory by quant and context

Quant4K ctx8K ctx32K ctx128K ctxMax tok/s (est., 8K)
2-bit (Q2_K) •1.7 GB Fits2.0 GB Fits3.9 GB Fits11.4 GB Fits262
3-bit (Q3_K_M) •2.0 GB Fits2.3 GB Fits4.2 GB Fits11.7 GB Fits205
4-bit (Q4_K_M) •2.2 GB Fits2.6 GB Fits4.4 GB Fits11.9 GB Fits166
5-bit (Q5_K_M) •2.5 GB Fits2.8 GB Fits4.7 GB Fits12.2 GB Fits142
6-bit (Q6_K) •2.7 GB Fits3.1 GB Fits4.9 GB Fits12.4 GB Fits123
8-bit (Q8_0) •3.3 GB Fits3.6 GB Fits5.5 GB Fits13.0 GB Fits95
16-bit (F16/BF16) •5.5 GB Fits5.8 GB Fits7.7 GB Fits15.2 GB Fits51

• real GGUF file size from ibm-granite/granite-3.3-2b-instruct-GGUF. Speed is a bandwidth ceiling (256 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

Granite 3.3 2B Instruct on every GPU and Mac · Every model that runs on the Ryzen AI Max+ 395 128GB · Other models that fit 64 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 (ibm-granite/granite-3.3-2b-instruct), AMD specs. Model data last checked 2026-11-18; refreshed hourly.