Files and memory at 8K context
| Quant | File | Files | + KV + runtime | Smallest fit | Source |
|---|---|---|---|---|---|
| Q4_K_M unsloth/GLM-4.7-Flash-GGUF | 17.05 GiB | 1 | unknown | 20 GB GPU (cache unknown) | HF files |
| Q5_K_M unsloth/GLM-4.7-Flash-GGUF | 19.94 GiB | 1 | unknown | 24 GB GPU (cache unknown) | HF files |
| Q8_0 unsloth/GLM-4.7-Flash-GGUF | 29.66 GiB | 1 | unknown | 32 GB GPU (cache unknown) | HF files |
| MLX 4-bit mlx-community/GLM-4.7-Flash-4bit | 15.70 GiB | 4 | unknown | 24 GB Mac (cache unknown) | HF files |
| MLX 8-bit mlx-community/GLM-4.7-Flash-8bit | 29.64 GiB | 7 | unknown | 48 GB Mac (cache unknown) | 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
Unknown
Arena text rating
Unknown
Scores describe the evaluated hosted variant, not each quantized file. AA and Arena are never averaged.
Reported speeds 7
| Hardware | Quant | Backend | Prompt t/s | Gen t/s | Date | Source |
|---|---|---|---|---|---|---|
| M5 Max (40c GPU) 64 GB | MLX 4bit | MLX | 2,093 | 71.9 | 2026-10-01 | omlx.ai |
| M5 Max (32c GPU) 36 GB | MLX 4bit | MLX | 1,696 | 71.6 | 2026-09-28 | omlx.ai |
| M5 Pro (20c GPU) 48 GB | MLX 4bit | MLX | 1,144 | 58 | 2026-09-25 | omlx.ai |
| M2 Max (30c GPU) 32 GB | MLX 4bit | MLX | 380 | 45.4 | 2026-08-20 | omlx.ai |
| M3 Max (30c GPU) 96 GB | MLX 4bit | MLX | 715 | 46.6 | 2026-08-19 | omlx.ai |
| M2 Max (38c GPU) 96 GB | MLX 4bit | MLX | 448 | 48.7 | 2026-08-11 | omlx.ai |
| M4 Pro (20c GPU) 48 GB | MLX 4bit | MLX | 561 | 51.4 | 2026-03-21 | omlx.ai |
Reported by the linked sources, not reproduced by LocalFit. All rows in the speed database
Community threads
- Quantized GLM-4.7-Flash with llama.cpp and opencode. Hugging Face · 2026-02-16
- Is FIM supported for code completion？ Hugging Face · 2026-02-12
Qualitative signals only; titles are not verified claims.