ibm-granite
granite 4.2 8b
granite 4.2 8b is a 8.79B dense granite model from ibm-granite. Configured context length is 131,072 tokens. License: apache-2.0. Quant sizes are the GGUF file bytes published on Hugging Face when a file exists.
Parameters8.8B
Context128K
Attentiongqa
Licenseapache-2.0
40 layers · 8 KV heads · head dim 128 · GGUF repo · config
Quant file sizes
| Quant | File size | Source |
|---|---|---|
| BF16 | 16.4 GB | Hugging Face file bytes |
| Q8_0 | 8.70 GB | Hugging Face file bytes |
| Q6_K | 6.72 GB | Hugging Face file bytes |
| Q5_K_M | 5.82 GB | Hugging Face file bytes |
| Q5_K_S | 5.68 GB | Hugging Face file bytes |
| Q5_1 | 6.17 GB | Hugging Face file bytes |
| Q5_0 | 5.68 GB | Hugging Face file bytes |
| Q4_K_M | 4.98 GB | Hugging Face file bytes |
| Q4_K_S | 4.74 GB | Hugging Face file bytes |
| Q4_1 | 5.20 GB | Hugging Face file bytes |
| Q4_0 | 4.71 GB | Hugging Face file bytes |
| Q3_K_L | 4.38 GB | Hugging Face file bytes |
| Q3_K_M | 4.05 GB | Hugging Face file bytes |
| Q3_K_S | 3.67 GB | Hugging Face file bytes |
| Q2_K | 3.18 GB | Hugging Face file bytes |
On common machines
| Device | Verdict | Quant | Speed |
|---|---|---|---|
| GeForce RTX 4060 Ti 8GB | Tight fit | Q5_K_M | 23–31 |
| GeForce RTX 3060 12GB | Tight fit | Q8_0 | 19–25 |
| GeForce RTX 4060 Ti 16GB | Runs | Q8_0 | 15–20 |
| GeForce RTX 4090 | Runs | BF16 | 29–40 |
| GeForce RTX 5090 | Runs | BF16 | 52–71 |
| Apple M4 Pro | Tight fit | BF16 | 8.2–11 |
| Apple M5 | Tight fit | Q8_0 | 7.9–11 |
| CPU only · DDR5-5600 | Runs on CPU | BF16 | 2.6–3.5 |
Smallest devices that fit it fully
- GeForce RTX 3050 6GB — Tight fit, Q3_K_M, 18–25 tok/s
- GeForce RTX 2060 6GB — Tight fit, Q3_K_M, 37–50 tok/s
- GeForce GTX 1660 Super — Tight fit, Q3_K_M, 37–50 tok/s
- GeForce GTX 1660 Ti — Tight fit, Q3_K_M, 32–43 tok/s
- Intel Arc A380 — Tight fit, Q3_K_M, 20–28 tok/s
- GeForce RTX 5060 Ti 8GB — Tight fit, Q5_K_M, 36–49 tok/s
Run it
Commands for the GeForce RTX 4060 Ti 16GB estimate (Q8_0).
llama.cpp
llama-server -hf ibm-granite/granite-4.2-8b-GGUF:Q8_0 -c 8192 -ngl 40
No matching name in the Ollama library snapshot.
LM Studio
lms get ibm-granite/granite-4.2-8b-GGUF # Then load quant Q8_0