Mac Studio M3 Ultra 512GB
$23,000
Parts
A complete machine needs no part checks.
| Component | Selection | Price | Where | Buy |
|---|---|---|---|---|
| Machine | Mac Studio M3 Ultra 512GB 512 GB unified · 819 GB/s | $23,000 | eBay | Buy |
| Total | $23,000 | |||
What this build runs
The most precise quant that fits, at 8k context. Estimates, not measurements.
| Model | Quant | Fits | Memory | Max context | Decode @8k | Prefill | Good for |
|---|---|---|---|---|---|---|---|
| DeepSeek V4 Flash DeepSeek | MXFP4 | fits | 146.6 GB | 1024k | 27tok/sest. | 175 tok/s | chat, coding |
| GLM 5.3 Flash Z.ai | Q8_0 | fits | 318.4 GB | 1024k | 12tok/sest. | 178 tok/s | – |
| Qwen 3.8 Flash Next Alibaba Qwen | Q8_0 | fits | 176.2 GB | 256k | 5tok/sest. | 90 tok/s | – |
| MiniMax M2.7 MiniMax | Q8_0 | fits | 229.0 GB | 200k | 23tok/sest. | 396 tok/s | chat |
| Qwen 3.8 27B Alibaba Qwen | Q8_0 | fits | 28.3 GB | 256k | 18tok/sest. | 196 tok/s | chat |
| Qwen 3.5 122B-A10B Alibaba Qwen | Q8_0 | fits | 121.9 GB | 256k | 30tok/sest. | 518 tok/s | chat, coding |
| Qwen 3.6 27B Alibaba Qwen | Q8_0 | fits | 27.9 GB | 256k | 18tok/sest. | 196 tok/s | chat |
| gpt-oss 120B OpenAI | MXFP4 | fits | 60.0 GB | 128k | 93tok/sest. | 879 tok/s | chat, coding, agents, long documents |
| Qwen3 Coder Next Alibaba Qwen | Q8_0 | fits | 79.9 GB | 256k | 73tok/sest. | 1309 tok/s | chat, coding, agents, long documents |
| Nemotron 3 Super 120B-A12B NVIDIA | Q8_0 | fits | 120.5 GB | 256k | 25tok/sest. | 457 tok/s | chat, coding, long documents |
| Gemma 4 31B | Q8_0 | fits | 32.4 GB | 256k | 16tok/sest. | 171 tok/s | chat |
| Mistral Small 4 119B Mistral AI | Q8_0 | fits | 118.6 GB | 1024k | 42tok/sest. | 677 tok/s | chat, coding |
| Qwen 3.6 35B-A3B Alibaba Qwen | Q8_0 | fits | 35.2 GB | 256k | 97tok/sest. | 1679 tok/s | chat, coding, agents, long documents |
| Qwen 3.5 27B Alibaba Qwen | Q8_0 | fits | 27.9 GB | 256k | 18tok/sest. | 196 tok/s | chat |
| Gemma 4 26B-A4B | Q8_0 | fits | 26.0 GB | 256k | 70tok/sest. | 1279 tok/s | chat, coding, agents, long documents |
| GLM 4.5 Air Z.ai | Q8_0 | fits | 111.5 GB | 128k | 16tok/sest. | 258 tok/s | chat |
| Qwen 3.5 35B-A3B Alibaba Qwen | Q8_0 | fits | 35.2 GB | 256k | 97tok/sest. | 1679 tok/s | chat, coding, agents, long documents |
| Qwen3 Next 80B-A3B Alibaba Qwen | Q8_0 | fits | 79.9 GB | 256k | 94tok/sest. | 1646 tok/s | chat, coding, agents, long documents |
| GLM 4.7 Flash Z.ai | Q8_0 | fits | 30.7 GB | 198k | 56tok/sest. | 940 tok/s | chat, coding, agents, long documents |
| Devstral Small 2 24B Mistral AI | Q8_0 | fits | 25.2 GB | 384k | 20tok/sest. | 221 tok/s | chat |
| Nemotron 3.5 Lightning 30B-A3B NVIDIA | UD-Q6_K_XL | fits | 33.3 GB | 256k | 93tok/sest. | 1749 tok/s | chat, coding, agents, long documents |
| gpt-oss 20B OpenAI | MXFP4 | fits | 12.1 GB | 128k | 121tok/sest. | 1215 tok/s | chat, coding, agents, long documents |
| Kimi Linear 48B-A3B Moonshot AI | Q8_0 | fits | 49.4 GB | 1024k | 97tok/sest. | 1596 tok/s | chat, coding, long documents |
| Qwen3 Coder 30B-A3B Alibaba Qwen | Q8_0 | fits | 31.6 GB | 256k | 80tok/sest. | 1215 tok/s | chat, coding, agents |
| Qwen3 32B Alibaba Qwen | Q8_0 | fits | 35.1 GB | 40k | 14tok/sest. | 151 tok/s | – |
| Gemma 4 12B | Q8_0 | fits | 12.9 GB | 256k | 41tok/sest. | 446 tok/s | chat, coding, long documents |
| Qwen 3.5 9B Alibaba Qwen | Q8_0 | fits | 9.8 GB | 256k | 54tok/sest. | 564 tok/s | chat, coding, long documents |
| Mistral Small 3.2 24B Mistral AI | Q8_0 | fits | 25.2 GB | 128k | 20tok/sest. | 221 tok/s | chat |
| Gemma 3 27B | Q8_0 | fits | 28.4 GB | 128k | 18tok/sest. | 199 tok/s | chat |
| DeepSeek R1 Distill Qwen 32B DeepSeek | Q8_0 | fits | 35.1 GB | 128k | 14tok/sest. | 158 tok/s | – |
| Llama 3.3 70B Meta | Q8_0 | fits | 73.0 GB | 128k | 7tok/sest. | 74 tok/s | – |
| Ministral 3 14B Mistral AI | Q8_0 | fits | 15.3 GB | 256k | 34tok/sest. | 366 tok/s | chat, coding |
| Qwen3 14B Alibaba Qwen | Q8_0 | fits | 16.5 GB | 40k | 31tok/sest. | 341 tok/s | chat, coding |
| DeepSeek R1 0528 Qwen3 8B DeepSeek | Q8_0 | fits | 9.9 GB | 128k | 54tok/sest. | 596 tok/s | chat, coding |
| Qwen3 8B Alibaba Qwen | Q8_0 | fits | 9.9 GB | 40k | 54tok/sest. | 596 tok/s | chat, coding |
| Phi-4 14B Microsoft | Q8_0 | fits | 16.7 GB | 16k | 31tok/sest. | 343 tok/s | chat, coding |
| Qwen 3.5 4B Alibaba Qwen | Q8_0 | fits | 5.1 GB | 256k | 111tok/sest. | 1136 tok/s | chat, coding, long documents |
| Ministral 3 8B Mistral AI | Q8_0 | fits | 10.1 GB | 256k | 52tok/sest. | 557 tok/s | chat, coding |
| Granite 4.2 8B IBM | Q8_0 | fits | 10.6 GB | 128k | 50tok/sest. | 553 tok/s | chat, coding |
| Gemma 4 E4B | Q8_0 | fits | 8.5 GB | 128k | 63tok/sest. | 684 tok/s | chat, coding, long documents |
| LFM2.5 8B-A1B Liquid AI | Q8_0 | fits | 9.4 GB | 125k | 218tok/sest. | 3992 tok/s | chat, coding, long documents |
| Llama 3.1 8B Meta | Q6_K | fits | 7.8 GB | 128k | 69tok/sest. | 615 tok/s | chat, coding |
| Qwen2.5 0.5B Alibaba Qwen | Q8_0 | fits | 1.2 GB | 32k | 843tok/sest. | 8356 tok/s | chat, coding, long documents |
| Qwen3.5 0.8B Alibaba Qwen | Q8_0 | fits | 1.5 GB | 256k | 572tok/sest. | 5749 tok/s | chat, coding, agents, long documents |
| Qwen3.5 2B Alibaba Qwen | Q5_K_M | fits | 2.1 GB | 256k | 342tok/sest. | 2358 tok/s | chat, coding, agents, long documents |
| DeepSeek R1 Distill Qwen 1.5B DeepSeek | Q8_0 | fits | 2.6 GB | 128k | 250tok/sest. | 2630 tok/s | chat, coding, long documents |
| MiniCPM5 2B openbmb | Q8_0 | fits | 3.5 GB | 128k | 176tok/sest. | 1738 tok/s | chat, coding, agents, long documents |
| LFM2.5 2.6B Liquid AI | Q8_0 | fits | 3.8 GB | 128k | 158tok/sest. | 1750 tok/s | chat, coding, agents, long documents |
| LFM2.5 VL 3B Liquid AI | Q8_0 | fits | 3.8 GB | 32k | 158tok/sest. | 1544 tok/s | chat, coding, agents, long documents |
| granite vision 4.1 4b IBM | Q8_0 | fits | 4.6 GB | 128k | 124tok/sest. | 1158 tok/s | chat, coding, long documents |
| granite 4.2 3b IBM | Q8_0 | fits | 4.9 GB | 128k | 117tok/sest. | 1245 tok/s | chat, coding, agents, long documents |
| LFM2.5 24B A2B Liquid AI | Q8_0 | fits | 25.0 GB | 128k | 114tok/sest. | 2097 tok/s | chat, coding, agents, long documents |
| Mellum2.1 12B A2.5B Thinking JetBrains | Q8_0 | fits | 12.8 GB | 128k | 113tok/sest. | 1924 tok/s | chat, coding, agents, long documents |
| Kimi VL A3B Moonshot AI | mlx-8bit | fits | 17.5 GB | 128k | 91tok/sest. | 1622 tok/s | chat, coding, agents, long documents |
| DeepSeek R1 Distill Qwen 7B DeepSeek | Q8_0 | fits | 8.6 GB | 128k | 62tok/sest. | 664 tok/s | chat, coding |
| DeepSeek R1 Distill Llama 8B DeepSeek | Q8_0 | fits | 9.6 GB | 128k | 55tok/sest. | 615 tok/s | chat, coding |
| Ministral 3 8B Reasoning 2512 Mistral AI | Q8_0 | fits | 10.1 GB | 256k | 52tok/sest. | 557 tok/s | chat, coding |
| GLM 4.6V Flash Z.ai | Q8_0 | fits | 10.3 GB | 128k | 52tok/sest. | 481 tok/s | chat, coding |
| GLM 4.1V 9B Thinking Z.ai | mlx-8bit | fits | 11.9 GB | 64k | 44tok/sest. | 481 tok/s | chat, coding |
| DeepSeek R1 Distill Qwen 14B DeepSeek | Q8_0 | fits | 16.8 GB | 128k | 31tok/sest. | 334 tok/s | chat, coding |
| MiniMax M2 MiniMax | Q8_0 | fits | 229.0 GB | 192k | 23tok/sest. | 396 tok/s | chat |
| MiniMax M2.1 MiniMax | Q8_0 | fits | 229.0 GB | 192k | 23tok/sest. | 396 tok/s | chat |
| MiniMax M2.5 MiniMax | Q8_0 | fits | 229.0 GB | 192k | 23tok/sest. | 395 tok/s | chat |
| Devstral Small 2507 Mistral AI | Q8_0 | fits | 25.2 GB | 128k | 20tok/sest. | 225 tok/s | chat |
| Magistral Small 2506 Mistral AI | Q8_0 | fits | 25.2 GB | 40k | 20tok/sest. | 225 tok/s | chat |
| Magistral Small 2507 Mistral AI | Q8_0 | fits | 25.2 GB | 40k | 20tok/sest. | 225 tok/s | chat |
| Magistral Small 2509 Mistral AI | Q8_0 | fits | 25.2 GB | 128k | 20tok/sest. | 221 tok/s | chat |
| GLM 4.5V Z.ai | Q8_0 | fits | 107.9 GB | 64k | 19tok/sest. | 296 tok/s | chat |
| GLM 4.6V Z.ai | Q8_0 | fits | 107.9 GB | 128k | 19tok/sest. | 296 tok/s | chat |
| DeepSeek V4 Flash 0731 DeepSeek | MXFP4 | fits | 147.0 GB | 1024k | 18tok/sest. | 124 tok/s | chat |
| DeepSeek V4 Flash Vision Exp DeepSeek | MXFP4 | fits | 155.7 GB | 1024k | 17tok/sest. | 123 tok/s | chat |
| granite 4.2 30b IBM | Q8_0 | fits | 31.6 GB | 128k | 16tok/sest. | 178 tok/s | chat |
| MiniMax M3 MiniMax | UD-Q6_K | fits | 332.5 GB | 426k770k with q8 cache | 16tok/sest. | 207 tok/s | chat |
| DeepSeek R1 DeepSeek | Q4_K_M | fits | 377.8 GB | 92k160k with q8 cache | 14tok/sest. | 112 tok/s | – |
| DeepSeek R1 0528 DeepSeek | Q4_K_M | fits | 378.3 GB | 86k151k with q8 cache | 14tok/sest. | 112 tok/s | – |
| DeepSeek V3 0324 DeepSeek | Q4_K_M | fits | 377.8 GB | 92k160k with q8 cache | 14tok/sest. | 112 tok/s | – |
| DeepSeek V3.1 DeepSeek | Q4_K_M | fits | 378.7 GB | 80k140k with q8 cache | 14tok/sest. | 112 tok/s | – |
| DeepSeek V3.2 DeepSeek | Q4_K_M | fits | 378.7 GB | 80k140k with q8 cache | 13tok/sest. | 110 tok/s | – |
| GLM 4.7 Z.ai | Q8_0 | fits | 358.3 GB | 78k145k with q8 cache | 7tok/sest. | 116 tok/s | – |
| DeepSeek R1 Distill Llama 70B DeepSeek | Q8_0 | fits | 73.0 GB | 128k | 7tok/sest. | 74 tok/s | – |
| Devstral 2 123B Instruct 2512 Mistral AI | Q8_0 | fits | 127.1 GB | 256k | 4tok/sest. | 42 tok/s | – |
| Mistral Medium 3.5 128B Mistral AI | Q8_0 | fits | 127.1 GB | 256k | 4tok/sest. | 41 tok/s | – |
| GLM 5.1 Z.ai | MXFP4 | won't run | 421.3 GB | – | – | – | – |
| GLM 5.2 Z.ai | UD-Q4_K_M | won't run | 435.2 GB | – | – | – | – |
| GLM 5.3 Z.ai | Q8_0 | won't run | 747.7 GB | – | – | – | – |
| Kimi K2 Moonshot AI | Q4_K_M | won't run | 579.3 GB | – | – | – | – |
| Kimi K2 Instruct 0905 Moonshot AI | Q4_K_M | won't run | 579.8 GB | – | – | – | – |
| Kimi K2 Thinking Moonshot AI | Q4_K_M | won't run | 579.8 GB | – | – | – | – |
| Kimi K2.5 Moonshot AI | Q4_K_M | won't run | 579.8 GB | – | – | – | – |