NVIDIA Nemotron 3 Ultra 550B-A55B
Files and memory at 8K context
| Quant | File | Files | + KV + runtime | Smallest fit | Source |
|---|---|---|---|---|---|
| Q4_K_M DevQuasar/nvidia.NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16-GGUF | 321 GiB | 24 | 323 GiB+ | over 192 GB GPU | HF files |
| Q5_K_M DevQuasar/nvidia.NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16-GGUF | 369 GiB | 30 | 370 GiB+ | over 192 GB GPU | HF files |
| Q8_0 unsloth/NVIDIA-Nemotron-3-Ultra-550B-A55B-GGUF | 544 GiB | 14 | 545 GiB+ | over 192 GB GPU | HF files |
| BF16 unsloth/NVIDIA-Nemotron-3-Ultra-550B-A55B-GGUF | 1,024 GiB | 24 | 1,025 GiB+ | over 192 GB GPU | HF files |
| MLX 4-bit pipenetwork/NVIDIA-Nemotron-3-Ultra-550B-A55B-MLX-4bit | 288 GiB | 96 | 289 GiB+ | 512 GB Mac | HF files |
| MLX 8-bit pipenetwork/NVIDIA-Nemotron-3-Ultra-550B-A55B-MLX-8bit | 544 GiB | 193 | 545 GiB+ | over 512 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
22.9
Variant: Nemotron 3 Ultra 550B A55B (Reasoning) · observed 2026-10-17 · source
Rank #27 of 75 researched 2026 models.
Arena text rating
1,445
95% CI 1438–1451 · 11,567 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 1
| Hardware | Quant | Backend | Prompt t/s | Gen t/s | Date | Source |
|---|---|---|---|---|---|---|
| Apple M3 Ultra (80-core GPU), 512 GB unified memory | 4bit (Nemotron-3-Ultra-550B-A55B) | MLX (oMLX server) | 157 | 16.2 | 2026-06-08 | omlx.ai |
Reported by the linked sources, not reproduced by LocalFit. All rows in the speed database