Yes — with 183.6 GB to spare
Granite 4.1 8B at Q4_K_M fits your Instinct MI300X entirely on the GPU at 8K context, at an estimated 649 tokens per second. There is room for its full 128K window.
Fully on GPU
8K context
Q4_K_M · 4.9 GB
Apache 2.0
Released 29 Apr 2026
Matches the old Granite 4.0 32B MoE at a quarter of the size. Fast, Apache 2.0, no reasoning traces.
The VRAM budget
weights 4.9 GB
Weights 4.9 GB
KV cache @ 8K 1.25 GB
Runtime overhead 0.6 GB
Free 183.6 GB of 190.4 GB
Quantisation ladder
| Quant | Weights | Total @ 8K | Max context | Tok/s | Quality | Fit |
|---|---|---|---|---|---|---|
| F16 | 16.4 GB | 18.2 GB | 128K | 196 | Reference | Long context |
| Q8_0 | 8.7 GB | 10.5 GB | 128K | 369 | −0.1% ppl | Long context |
| Q6_K | 6.7 GB | 8.6 GB | 128K | 478 | −0.4% ppl | Long context |
| Q5_K_M | 5.8 GB | 7.7 GB | 128K | 553 | −0.8% ppl | Long context |
| Q4_K_M | 4.9 GB | 6.8 GB | 128K | 649 | −1.9% ppl | Recommended |
| Q3_K_M | 4.0 GB | 5.9 GB | 128K | 802 | −5.4% ppl | Long context |
| Q2_K | 3.4 GB | 5.3 GB | 128K | 936 | −15% ppl | Long context |
Quality is the published perplexity delta against f16 weights. Max context assumes an f16 KV cache; q8_0 roughly doubles it.
How to run it
$ ollama pull granite4.1:8b $ OLLAMA_CONTEXT_LENGTH=8192 \ ollama run granite4.1:8b
The default. One binary, a model registry, an OpenAI-compatible port. More on Ollama.
01Download is 4.9 GB. Keep it on an SSD — a first load off a spinning disk takes minutes.
02Close anything else holding VRAM. A browser with hardware acceleration can sit on 1–2 GB.
03There is room to go to the model's full 128K context on this card.