Apple M5 Max (18-core CPU / 32-core GPU): 36 GB, what it runs and what it costs
Specs checked · Aliteq Team · How we compute this
Some links on this page are referral links: we may earn a commission, at no cost to you.
The short answer
The Apple M5 Max (18-core CPU / 32-core GPU) has 36 GB of unified memory and 460 GB/s of memory bandwidth (2026). The CPU and the operating system share that memory, so a model gets somewhat less than the full amount. At Q4_K_M quality or better with an 8K context it runs 39 of the open models we track; the largest that fits with room to spare is Qwen3.5 35B-A3B (MoE). We don't see it for rent on the providers we track right now.
- Unified memory
- 36 GB
- Bandwidth
- 460 GB/s
- Board power
- —
- Launched
- 2026
Source: Apple specification · bandwidth not yet confirmed by a second source, so we don't estimate speed from it
Buy it or rent it
Not for rent on the providers we track right now. See every card that is.
Which AI models run on the Apple M5 Max (18-core CPU / 32-core GPU)
Computed by our VRAM engine for an 8K context: the best quality each model fits at in 36 GB. Speeds are upper-bound estimates from memory bandwidth, not benchmarks.
| Model | Best fit | VRAM used | ~tok/s |
|---|---|---|---|
| Mixtral 8x7B (MoE) | Q5_K_M | 33 GB91% · tight | — |
| Qwen3.5 35B-A3B (MoE) | Q6_K | 28 GB79% | — |
| Qwen3.6 35B-A3B (MoE) | Q6_K | 28 GB79% | — |
| Ornith-1.5 35B-A3B (MoE) | Q6_K | 28 GB79% | — |
| DeepSeek R1 Distill 32B | Q8_0 | 35 GB98% · tight | — |
| Qwen2.5 32B Instruct | Q8_0 | 35 GB98% · tight | — |
| Qwen2.5 Coder 32B | Q8_0 | 35 GB98% · tight | — |
| Qwen3 32B | Q8_0 | 35 GB98% · tight | — |
| NVIDIA Nemotron 3 Nano 30B-A3B (MoE) | Q8_0 | 32 GB89% | — |
| Gemma 4 31B | Q8_0 | 32 GB90% | — |
| GLM-4.7-Flash (MoE) | Q8_0 | 32 GB89% | — |
| Qwen3-VL 30B-A3B Instruct (MoE) | Q8_0 | 32 GB90% | — |
Fits only below Q4 quality: Qwen3-Next 80B-A3B (MoE) (Q2_K), Qwen3-Coder-Next 80B-A3B (MoE) (Q2_K), Llama 3.3 70B Instruct (Q3_K_M), DeepSeek R1 Distill 70B (Q3_K_M).
Too big for 36 GB (rent a bigger card instead): what it costs to run GLM-4.5-Air (MoE), what it costs to run Llama 4 Scout 17B-16E (MoE), what it costs to run gpt-oss 120B (MoE), what it costs to run Mistral Small 4 119B (MoE).
Check a model and context of your own in the VRAM calculator.
Our guides on the Apple M5 Max (18-core CPU / 32-core GPU)
Cards to compare it with
- Apple M4 Max (14-core CPU / 32-core GPU) · 36 GB
- RTX 5000 Ada Generation · 32 GB
- RTX 5090 · 32 GB
- RTX A6000 · 48 GB
- RTX A5000 · 24 GB
- AMD Radeon PRO W7900 · 48 GB
Every card by VRAM tier: all GPUs for gaming and local AI.