Can MacBook Neo A18 Pro 8GB run Llama-3.2-3B-Instruct?

Apple · 8 GB LPDDR5X · 60 GB/s bandwidth · Meta · 3.2B · model ctx up to 128K

Fit & speed by quant and context

Fits = weights + KV(ctx) + 0.5 GB overhead ≤ 95% of VRAM. Measured anchors are context-agnostic; the fit verdict is recomputed per context. A missing tok/s means the model is far beyond this card (offload-only territory).

VRAM breakdown at 8K context

Other GPUs that run Llama-3.2-3B-Instruct

• B200 192GB SXM — ≈2349.1 tok/s Q4 est.
• MI300X 192GB — ≈1535.9 tok/s Q4 est.
• H200 141GB SXM5 — ≈1409.5 tok/s Q4 est.
• H100 80GB SXM5 — ≈983.7 tok/s Q4 est.
• Gaudi 3 128GB — ≈949.3 tok/s Q4 est.
• A100 80GB SXM4 — ≈598.7 tok/s Q4 est.
• RTX 5090 32GB — ≈526.2 tok/s Q4 est.
• RTX 4090 24GB — ≈269.0 tok/s Q4 est.

All GPUs for Llama-3.2-3B-Instruct →

Other models for the MacBook Neo A18 Pro 8GB

• MiniCPM5-1B — ≈70.0 tok/s Q4 est.
• Qwen3.5-2B — ≈37.8 tok/s Q4 est.
• Nanbeige4.2-3B — ≈19.2 tok/s Q4 est.
• Qwen3-4B-Instruct-2507 — ≈18.9 tok/s Q4 est.
• NVIDIA-Nemotron-3-Nano-4B-BF16 — ≈18.9 tok/s Q4 est.
• Qwen3-4B-Thinking-2507 — ≈18.9 tok/s Q4 est.
• gemma-4-E2B-it — ≈14.8 tok/s Q4 est.
• Mistral-7B-Instruct-v0.3 — ≈10.5 tok/s Q4 est.

All models on the MacBook Neo A18 Pro 8GB →