Granite 4.2 8B
Files and memory at 8K context
| Quant | File | Files | + KV + runtime | Smallest fit | Source |
|---|---|---|---|---|---|
| Q4_K_M ibm-granite/granite-4.2-8b-GGUF | 4.98 GiB | 1 | 7.2 GiB | 8 GB GPU | HF files |
| Q5_K_M ibm-granite/granite-4.2-8b-GGUF | 5.82 GiB | 1 | 8.1 GiB | 12 GB GPU | HF files |
| Q8_0 ibm-granite/granite-4.2-8b-GGUF | 8.70 GiB | 1 | 11.0 GiB | 12 GB GPU | HF files |
| FP16 mradermacher/granite-4.2-8b-GGUF | 16.38 GiB | 1 | 18.6 GiB | 20 GB GPU | HF files |
| MLX 4-bit lmstudio-community/granite-4.2-8b-MLX-4bit | 4.61 GiB | 1 | 6.9 GiB | 16 GB Mac | HF files |
| MLX 8-bit lmstudio-community/granite-4.2-8b-MLX-8bit | 8.70 GiB | 2 | 10.9 GiB | 16 GB Mac | HF files |
Sizes are exact file bytes (text weights; projectors listed separately). “+” = recurrent state or other terms unknown. Smallest fit adds 0.5 GiB spare; Macs use 75% of memory.
Quality
Artificial Analysis Intelligence Index
11.1
Variant: Granite 4.2 8B · observed 2026-10-17 · source
Rank #56 of 75 researched 2026 models.
Arena text rating
1,320
95% CI 1309–1331 · 3,254 votes · published 2026-10-15 · dataset
Exact listed model name; hosted reasoning/quantization may differ from local checkpoint.
Scores describe the evaluated hosted variant, not each quantized file. AA and Arena are never averaged.
Reported speeds 0
No reported measurement yet. Submit one.
Community threads
- Reasoning support not detected by llama.cpp Hugging Face · 2026-09-13
Qualitative signals only; titles are not verified claims.