G9v3-39A5B (preview)
Files and memory at 8K context
| Quant | File | Files | + KV + runtime | Smallest fit | Source |
|---|---|---|---|---|---|
| Q4_K_M linuxid10t/G9v3-39A5B-GGUF | 21.94 GiB | 1 | 23.2 GiB | 24 GB GPU | HF files |
| Q8_0 linuxid10t/G9v3-39A5B-GGUF | 38.63 GiB | 1 | 39.9 GiB | 48 GB GPU | HF files |
| MLX 4-bit mixed precision rapid-mlx/G9v3-39A5B-MLX-4bit | 21.24 GiB | 5 | 22.5 GiB | 32 GB Mac | HF files |
No verified file: Q5_K_M, FP16, MLX 8-bit. 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
21.8 AA estimate
Variant: G9v3-39A5B · observed 2026-10-17 · source
Rank #29 of 75 researched 2026 models.
Arena text rating
Unknown
Scores describe the evaluated hosted variant, not each quantized file. AA and Arena are never averaged.
Reported speeds 1
| Hardware | Quant | Backend | Prompt t/s | Gen t/s | Date | Source |
|---|---|---|---|---|---|---|
| M5 Pro (20c GPU) 48 GB | MLX 4bit | MLX | 263 | 62.9 | 2026-08-06 | omlx.ai |
Reported by the linked sources, not reproduced by LocalFit. All rows in the speed database
Community threads
- GGUF Version? Hugging Face · 2026-08-04
- Is model MTP trained? Hugging Face · 2026-08-20
Qualitative signals only; titles are not verified claims.