Mac Studio M3 Ultra 96GB
96 GB unified · 819 GB/s
Prices
Every retailer we track in India, cheapest first. Buy links may earn us a commission.
No retailer in this market lists it right now.
What it runs
Every model in the catalogue, with the best quant for this hardware. Estimates, not measurements.
| Model | Quant | Fits | Memory | Max context | Decode @64k | Prefill | Good for |
|---|---|---|---|---|---|---|---|
| Qwen 3.8 27B Alibaba Qwen | Q8_0 | fits | 32.2 GB | 256k | 16tok/sest. | 164 tok/s | chat |
| Qwen 3.5 122B-A10B Alibaba Qwen | Q3_K_M | fits | 55.2 GB | 256k | 51tok/sest. | 341 tok/s | chat, coding |
| Qwen 3.6 27B Alibaba Qwen | Q8_0 | fits | 31.7 GB | 256k | 16tok/sest. | 164 tok/s | chat |
| Qwen3 Coder Next Alibaba Qwen | Q6_K | fits | 63.7 GB | 256k | 66tok/sest. | 789 tok/s | chat, coding, agents, long documents |
| gpt-oss 120B OpenAI | MXFP4 | fits | 62.3 GB | 128k | 58tok/sest. | 528 tok/s | chat, coding, agents, long documents |
| Nemotron 3 Super 120B-A12B NVIDIA | UD-Q3_K_M | fits | 59.1 GB | 256k | 48tok/sest. | 396 tok/s | chat, coding, long documents |
| Gemma 4 31B | Q8_0 | fits | 37.2 GB | 256k | 14tok/sest. | 149 tok/s | chat |
| Mistral Small 4 119B Mistral AI | UD-Q4_K_M | fits | 71.1 GB | 97k154k with q8 cache | 54tok/sest. | 335 tok/s | chat, coding |
| Qwen 3.6 35B-A3B Alibaba Qwen | Q8_0 | fits | 36.7 GB | 256k | 71tok/sest. | 985 tok/s | chat, coding, agents, long documents |
| Qwen 3.5 27B Alibaba Qwen | Q8_0 | fits | 31.7 GB | 256k | 16tok/sest. | 164 tok/s | chat |
| Gemma 4 26B-A4B | Q8_0 | fits | 27.4 GB | 256k | 56tok/sest. | 1009 tok/s | chat, coding, agents, long documents |
| GLM 4.5 Air Z.ai | Q3_K_M | fits | 65.8 GB | 97k128k with q8 cache | 15tok/sest. | 104 tok/s | chat, coding |
| Qwen3 Next 80B-A3B Alibaba Qwen | Q6_K | fits | 63.6 GB | 256k | 78tok/sest. | 900 tok/s | chat, coding, agents, long documents |
| Qwen 3.5 35B-A3B Alibaba Qwen | Q8_0 | fits | 36.7 GB | 256k | 71tok/sest. | 985 tok/s | chat, coding, agents, long documents |
| GLM 4.7 Flash Z.ai | Q8_0 | fits | 33.9 GB | 198k | 36tok/sest. | 489 tok/s | chat, coding, agents, long documents |
| Devstral Small 2 24B Mistral AI | Q8_0 | fits | 34.3 GB | 296k384k with q8 cache | 15tok/sest. | 161 tok/s | chat |
| Nemotron 3.5 Lightning 30B-A3B NVIDIA | UD-Q6_K_XL | fits | 34.0 GB | 256k | 85tok/sest. | 1215 tok/s | chat, coding, agents, long documents |
| Kimi Linear 48B-A3B Moonshot AI | Q8_0 | fits | 50.2 GB | 1024k | 85tok/sest. | 791 tok/s | chat, coding, long documents |
| gpt-oss 20B OpenAI | MXFP4 | fits | 13.8 GB | 128k | 79tok/sest. | 754 tok/s | chat, coding, agents, long documents |
| Qwen3 Coder 30B-A3B Alibaba Qwen | Q8_0 | fits | 37.2 GB | 256k | 33tok/sest. | 353 tok/s | chat, coding, agents |
| Qwen3 32B Alibaba Qwen | Q8_0 | fits | 49.4 GB | 40k | 10tok/sest. | 83 tok/s | – |
| Gemma 4 12B | Q8_0 | fits | 14.1 GB | 256k | 38tok/sest. | 388 tok/s | chat, coding, long documents |
| Qwen 3.5 9B Alibaba Qwen | Q8_0 | fits | 11.9 GB | 256k | 45tok/sest. | 475 tok/s | chat, coding, long documents |
| Gemma 3 27B | Q8_0 | fits | 33.1 GB | 128k | 15tok/sest. | 183 tok/s | chat |
| Mistral Small 3.2 24B Mistral AI | Q8_0 | fits | 34.3 GB | 128k | 15tok/sest. | 161 tok/s | chat |
| DeepSeek R1 Distill Qwen 32B DeepSeek | Q8_0 | fits | 49.4 GB | 128k | 10tok/sest. | 103 tok/s | – |
| Llama 3.3 70B Meta | Q5_K_M | fits | 67.5 GB | 78k128k with q8 cache | 8tok/sest. | 49 tok/s | – |
| Ministral 3 14B Mistral AI | Q8_0 | fits | 24.3 GB | 256k | 21tok/sest. | 227 tok/s | chat, coding |
| Qwen3 14B Alibaba Qwen | Q8_0 | fits | 25.6 GB | 40k | 20tok/sest. | 199 tok/s | chat, coding |
| DeepSeek R1 0528 Qwen3 8B DeepSeek | Q8_0 | fits | 18.1 GB | 128k | 29tok/sest. | 314 tok/s | chat, coding |
| Qwen3 8B Alibaba Qwen | Q8_0 | fits | 18.1 GB | 40k | 29tok/sest. | 314 tok/s | chat, coding |
| Phi-4 14B Microsoft | Q8_0 | fits | 28.0 GB | 16k | 18tok/sest. | 199 tok/s | chat, coding |
| Qwen 3.5 4B Alibaba Qwen | Q8_0 | fits | 7.2 GB | 256k | 80tok/sest. | 823 tok/s | chat, coding, long documents |
| Ministral 3 8B Mistral AI | Q8_0 | fits | 17.9 GB | 256k | 29tok/sest. | 310 tok/s | chat, coding |
| Granite 4.2 8B IBM | Q8_0 | fits | 19.7 GB | 128k | 27tok/sest. | 287 tok/s | chat, coding |
| Gemma 4 E4B | Q8_0 | fits | 10.4 GB | 128k | 53tok/sest. | 622 tok/s | chat, coding, long documents |
| LFM2.5 8B-A1B Liquid AI | Q8_0 | fits | 12.4 GB | 125k | 75tok/sest. | 1327 tok/s | chat, coding, long documents |
| Llama 3.1 8B Meta | Q6_K | fits | 15.1 GB | 128k | 35tok/sest. | 337 tok/s | chat, coding |
| DeepSeek V4 Flash DeepSeek | UD-Q2_K_XL | won't run | 96.5 GB | – | – | – | – |
| GLM 5.3 Flash Z.ai | UD-Q2_K_XL | won't run | 103.0 GB | – | – | – | – |
| Qwen 3.8 Flash Next Alibaba Qwen | UD-Q2_K_XL | won't run | 76.0 GB | – | – | – | – |
| MiniMax M2.7 MiniMax | UD-Q2_K_XL | won't run | 86.6 GB | 5k9k with q8 cache | – | – | – |
Specifications
From the manufacturer's published specifications; sources are linked where we have them.
- Chip
- Apple M3 Ultra (80-core GPU)
- Memory bus
- 1024-bit
- Memory bandwidth
- 819 GB/s
- FP16 compute
- 28 TFLOPS
- Software
- Metal
- Unified memory
- 96 GB
- GPU can use by default
- 72 GB
- GPU can use, tuned
- 88 GB
- Operating system
- macos
- Processor
- M3 Ultra 32-core
- Storage
- 1000 GB