Apple M2 (8GB Unified)
8 GB Unified · 100 GB/s
From
$899
Estimated street price
VRAM
8 GB
Bandwidth
100 GB/s
TDP
22W
Models
24
Tier
Limited
The Apple M2 (8GB Unified) with 8 GB unified memory can handle 24 AI models across chat, coding, ai_coding. Best performance: Llama 3.2 1B Instruct at 22 tok/s (good). Current price: approximately $899.
Source: OwnRig methodology
8 GB
100 GB/s
Unified
22W
10
MacBook Air 13" (2022), MacBook Air 15" (2023), MacBook Pro 13" (2022)
Builder Capability: Limited
Insufficient VRAM for most AI coding workflows.
Inference Backends
The software stacks that matter most for real-world inference on this device.
Metal
productionPrimary backend for Apple Silicon. Only 5–6GB available for models after macOS overhead — limits to small models.
What it can run
24 models| Arcee Trinity Mini 26B | Q3_K_M | – | Not viable |
| Arcee Trinity Nano 6B | Q5_K_M | 18 tok/s | Good |
| DeepSeek V3 | Q2_K | – | Not viable |
| Gemma 3 27B | Q4_K_M | – | Not viable |
| Gemma 3 4B | Q8_0 | 6 tok/s | Marginal |
| Gemma 4 26B-A4B | Q3_K_M | – | Not viable |
| Gemma 4 31B | Q3_K_M | – | Not viable |
| Gemma 4 E2B | Q8_0 | 9 tok/s | Acceptable |
| Gemma 4 E4B | Q8_0 | – | Not viable |
| GigaChat Lightning 10B | Q8_0 | – | Not viable |
| Llama 3.1 8B Instruct | Q4_K_M | 14 tok/s | Good |
| Llama 3.2 11B Vision | Q8_0 | – | Not viable |
| Llama 3.2 1B Instruct | Q8_0 | 22 tok/s | Good |
| Llama 3.2 3B Instruct | Q8_0 | 14 tok/s | Good |
| NVIDIA Nemotron-3-super-120B-A12B | Q2_K | – | Not viable |
| Phi-4 Mini | Q8_0 | 14 tok/s | Good |
| Qwen 2.5 Coder 32B Instruct | Q4_K_M | – | Not viable |
| Qwen 2.5 Coder 7B Instruct | Q5_K_M | 9 tok/s | Acceptable |
| Qwen3.5-122B-A10B | Q3_K_M | – | Not viable |
| Qwen3.5-27B | Q3_K_M | – | Not viable |
| Qwen3.5-397B (MoE) | Q2_K | – | Not viable |
| Qwen3.6-27B | Q3_K_M | – | Not viable |
| Stable Diffusion 3.5 Large | FP16 | – | Not viable |
| Whisper Large V3 Turbo | FP16 | – | Good |
Showing 24 of 24 entries
Buy Used Mac
Prices and availability vary. Inspect hardware before purchasing. Some links may be affiliate links.
Frequently Asked Questions
- What AI models can Apple M2 (8GB Unified) run?
- The Apple M2 (8GB Unified) can run 24 AI models. Top performers include Llama 3.2 1B Instruct, Arcee Trinity Nano 6B, Llama 3.1 8B Instruct. See the full compatibility table above for speeds and quality ratings.
- Is Apple M2 (8GB Unified) good for AI coding?
- With 8 GB, the Apple M2 (8GB Unified) has limited VRAM for AI coding workflows.
- How much VRAM does Apple M2 (8GB Unified) have?
- The Apple M2 (8GB Unified) has 8 GB of unified memory with 100 GB/s bandwidth.
- Can Apple M2 (8GB Unified) run 70B models?
- 70B models can run on the Apple M2 (8GB Unified) with CPU offloading, but performance will be reduced. Consider a GPU with 48GB+ VRAM for full-speed 70B inference.
- Is Apple M2 (8GB Unified) worth it for AI?
- At $899, the Apple M2 (8GB Unified) offers 8 GB VRAM and runs 24 AI models. It works for smaller models and experimentation.
Own this GPU?
See every AI model it supports, expected performance, and how to build around it.