IBM Granite 4.2 30B
Files and memory at 8K context
| Quant | File | Files | + KV + runtime | Smallest fit | Source |
|---|---|---|---|---|---|
| Q4_K_M ibm-granite/granite-4.2-30b-GGUF | 16.50 GiB | 1 | 19.5 GiB | 24 GB GPU | HF files |
| Q5_K_M ibm-granite/granite-4.2-30b-GGUF | 19.35 GiB | 1 | 22.3 GiB | 24 GB GPU | HF files |
| Q8_0 ibm-granite/granite-4.2-30b-GGUF | 28.98 GiB | 1 | 32.0 GiB | 48 GB GPU | HF files |
| MLX 4-bit lmstudio-community/granite-4.2-30b-MLX-4bit | 15.34 GiB | 4 | 18.3 GiB | 32 GB Mac | HF files |
| MLX 8-bit lmstudio-community/granite-4.2-30b-MLX-8bit | 28.97 GiB | 6 | 32.0 GiB | 48 GB Mac | HF files |
No verified file: FP16. 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
12.7
Variant: Granite 4.2 30B · observed 2026-10-17 · source
Rank #50 of 75 researched 2026 models.
Arena text rating
1,361
95% CI 1351–1372 · 3,401 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 2
| Hardware | Quant | Backend | Prompt t/s | Gen t/s | Date | Source |
|---|---|---|---|---|---|---|
| M1 Max (32c GPU) 64 GB | MLX 4bit | MLX | 97.1 | 16.2 | 2026-10-08 | omlx.ai |
| M5 Max (40c GPU) 128 GB | MLX 4bit | MLX | 827 | 29.9 | 2026-08-27 | omlx.ai |
Reported by the linked sources, not reproduced by LocalFit. All rows in the speed database
Community threads
- Thinking dilemmas Hugging Face · 2026-09-03
- Thank you for the contribution Hugging Face · 2026-09-01
Qualitative signals only; titles are not verified claims.