M4 Ultra (192 GB)
8 records across 8 models (quants: Q4_K - Medium).
| hardware | model | quant | backend | tok/s | pp tok/s | W | ttft s | ctx/tg | date | class | flags | source |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| M4 Ultra (192 GB) | DeepSeek R1 70B | Q4_K - Medium | MLX | 21 | – | – | 2.2 | – | – | sourced | – | source |
| M4 Ultra (192 GB) | gpt-oss 120B | Q4_K - Medium | MLX | 13 | – | – | 4.2 | – | – | sourced | – | source |
| M4 Ultra (192 GB) | Llama 3.3 70B | Q4_K - Medium | MLX | 21 | – | – | 2 | – | – | sourced | – | source |
| M4 Ultra (192 GB) | Llama 4 Scout 17B-16E | Q4_K - Medium | MLX | 30 | – | – | 1.3 | – | – | sourced | – | source |
| M4 Ultra (192 GB) | Mistral Small 4 119B | Q4_K - Medium | MLX | 45 | – | – | 0.7 | – | – | sourced | – | source |
| M4 Ultra (192 GB) | Qwen 2.5 72B | Q4_K - Medium | Ollama | 21 | – | – | 2.5 | – | – | sourced | – | source |
| M4 Ultra (192 GB) | Qwen 3 235B-A22B | Q4_K - Medium | MLX | 22 | – | – | 1.5 | – | – | sourced | – | source |
| M4 Ultra (192 GB) | Qwen 3 32B | Q4_K - Medium | Ollama | 45 | – | – | 1 | – | – | sourced | – | source |
All rows cite their source. Provenance: community 40018 sourced 1085