M2 Max 96GB
Apple Silicon
96GB VRAM · 400 GB/s bandwidth · 27.2 FP16 TFLOPS · 60W TDP · $2,199 street price
The M2 Max 96GB has 96GB of VRAM with 400 GB/s memory bandwidth and 27.2 TFLOPS FP16 compute. At Q4 quantization, it can comfortably run Gemma 3 4B (84 tok/s), Qwen 2.5 7B (44 tok/s), Llama 3.1 8B (41 tok/s). Models larger than ~163B parameters won't fit even at Q4. Electricity cost is approximately $6/month at 60W TDP.
What LLMs can you run?
| Model | Params | Q4 Weight | Fit | Decode |
|---|---|---|---|---|
| Gemma 3 4B | 4.0B | 2 GB | comfortable | 84 tok/s |
| Qwen 2.5 7B | 7.6B | 4 GB | comfortable | 44 tok/s |
| Llama 3.1 8B | 8.0B | 4 GB | comfortable | 41 tok/s |
| Mistral Small 24B | 24.0B | 12 GB | tight | 14 tok/s |
| Gemma 3 27B | 27.4B | 14 GB | tight | 12 tok/s |
| Qwen 2.5 Coder 32B | 32.5B | 16 GB | tight | 10 tok/s |
| Llama 3.3 70B | 70.6B | 35 GB | tight | 4 tok/s |
| Qwen 2.5 72B | 72.7B | 36 GB | tight | 4 tok/s |
| Llama 3.1 405B | 405B | 202 GB | won't fit | |
| DeepSeek R1 671B | 671B | 336 GB | won't fit |
Similar GPUs
| GPU | VRAM | BW | TFLOPS | TDP |
|---|---|---|---|---|
| RTX PRO 6000 Blackwell Server | 96GB | 1790 GB/s | 126.0 | 600W |
| H100 PCIe 96 GB | 96GB | 3360 GB/s | 248.3 | 700W |
| H100 SXM5 96 GB | 96GB | 3360 GB/s | 267.6 | 700W |
| H100 SXM5 94 GB | 94GB | 3360 GB/s | 267.6 | 700W |
| H100 NVL 94 GB | 94GB | 3940 GB/s | 241.3 | 400W |
Compare with another GPU
Select another GPU to compare specs and model performance side by side.