Every memory option compared, 546 to 614 GB/s memory bandwidth. Updated 2 November 2026.
M4 Max / M5 Max are sold with 36, 48, 64, 128 GB of memory. Counting the 31 models tracked here: 36 GB runs 26 models well; 48 GB runs 27 models well; 64 GB runs 28 models well; 128 GB runs 30 models well. Memory bandwidth is 546 to 614 GB/s, about 4.5 times the memory speed of a base M4 (120 GB/s), so the same model generates words proportionally faster. Apple Silicon shares one pool of memory between the processor and the GPU, and macOS lets the GPU use about two thirds of it on Macs up to 36 GB and about three quarters above that, so the memory you buy decides which models load at all.
| Model | 36 GB | 48 GB | 64 GB | 128 GB |
|---|---|---|---|---|
| Qwen 3.5 122B A10B | No | No | No | Great 28/s |
| gpt-oss 120B | No | No | No | Great 53/s |
| Llama 3.3 70B | No | No | Good 7/s | Good 7/s |
| Wan 2.2 14B | Slow about 15 min | Great about 8 min | Great about 8 min | Great about 8 min |
| LTX-2.3 | Great about 1 min | Great about 1 min | Great about 1 min | Great about 1 min |
| Qwen 3.6 35B A3B | Tight 74/s | Great 82/s | Great 82/s | Great 82/s |
| Qwen 3.5 35B A3B | Great 86/s | Great 86/s | Great 86/s | Great 86/s |
| Gemma 4 31B | Great 14/s | Great 14/s | Great 14/s | Great 14/s |
| DeepSeek R1 32B | Great 14/s | Great 14/s | Great 14/s | Great 14/s |
| Qwen3 Coder 30B | Great 79/s | Great 79/s | Great 79/s | Great 79/s |
| Qwen 3.6 27B | Great 16/s | Great 16/s | Great 16/s | Great 16/s |
| Gemma 4 26B A4B | Great 61/s | Great 61/s | Great 61/s | Great 61/s |
| Z-Image Turbo 6B | Great about 15 s | Great about 15 s | Great about 15 s | Great about 15 s |
| Wan 2.2 5B | Great about 4 min | Great about 4 min | Great about 4 min | Great about 4 min |
| gpt-oss 20B | Great 65/s | Great 65/s | Great 65/s | Great 65/s |
| ACE-Step 1.5 | Great about 47 s | Great about 47 s | Great about 47 s | Great about 47 s |
| FLUX.2 klein 4B | Great about 11 s | Great about 11 s | Great about 11 s | Great about 11 s |
| Phi-4 14B | Great 32/s | Great 32/s | Great 32/s | Great 32/s |
| Gemma 4 12B | Great 36/s | Great 36/s | Great 36/s | Great 36/s |
| Stable Diffusion XL | Great about 7 s | Great about 7 s | Great about 7 s | Great about 7 s |
| Qwen3 TTS 1.7B | Great about 17 s | Great about 17 s | Great about 17 s | Great about 17 s |
| Gemma 4 E4B | Great 89/s | Great 89/s | Great 89/s | Great 89/s |
| Qwen 3.5 9B | Great 44/s | Great 44/s | Great 44/s | Great 44/s |
| Qwen3 VL 8B | Great 47/s | Great 47/s | Great 47/s | Great 47/s |
| DeepSeek R1 8B | Great 56/s | Great 56/s | Great 56/s | Great 56/s |
| Qwen2.5 Coder 7B | Great 61/s | Great 61/s | Great 61/s | Great 61/s |
| Gemma 4 E2B | Great 143/s | Great 143/s | Great 143/s | Great 143/s |
| Whisper large-v3 turbo | Great about 25 s | Great about 25 s | Great about 25 s | Great about 25 s |
| Qwen 3.5 4B | Great 87/s | Great 87/s | Great 87/s | Great 87/s |
| Kokoro TTS | Great about 4 s | Great about 4 s | Great about 4 s | Great about 4 s |
Text models show words a second. Image, video and voice models show the time per job (image, 4 second clip, minute of speech or similar).
As an Amazon Associate I earn from qualifying purchases.
About 34 GB is usable by the GPU. The largest chat or coding model that fits is Qwen 3.6 35B A3B (24.6 GB), generating about 74 words a second with 9.6 GB to spare. The quickest is Gemma 4 E2B at about 143 words a second. Closest miss: Llama 3.3 70B, which needs about 45.6 GB (11 GB more than this configuration can give the GPU).
Find M5 Max Macs with 36 GB on Amazon
About 46 GB is usable by the GPU. The largest chat or coding model that fits is Qwen 3.6 35B A3B (24.6 GB), generating about 82 words a second with 21 GB to spare. The quickest is Gemma 4 E2B at about 143 words a second. Closest miss: Llama 3.3 70B, which needs about 45.6 GB (1 GB more than this configuration can give the GPU).
Find M5 Max Macs with 48 GB on Amazon
About 61 GB is usable by the GPU. The largest chat or coding model that fits is Llama 3.3 70B (45.6 GB), generating about 7 words a second with 15.2 GB to spare. The quickest is Gemma 4 E2B at about 143 words a second. Closest miss: gpt-oss 120B, which needs about 68.8 GB (8 GB more than this configuration can give the GPU).
Find M5 Max Macs with 64 GB on Amazon
About 122 GB is usable by the GPU. The largest chat or coding model that fits is Qwen 3.5 122B A10B (85.5 GB), generating about 28 words a second with 36.1 GB to spare. The quickest is Gemma 4 E2B at about 143 words a second. Closest miss: DeepSeek V3 671B, which needs about 424.7 GB (303 GB more than this configuration can give the GPU).
For everyday chat, 36 GB runs 26 of 31 models well. 64 GB to 128 GB unlocks 2 more: Qwen 3.5 122B A10B, gpt-oss 120B.
Yes, from 36 GB of memory, at about 14 words a second.
36 GB runs gpt-oss 20B at about 65 words a second; 48 GB runs gpt-oss 20B at about 65 words a second; 64 GB runs gpt-oss 20B at about 65 words a second; 128 GB runs gpt-oss 20B at about 65 words a second.
You are already at the top end. Nothing popular is out of reach except the giant server models.