Can I run Llama-3.2-3B-Instruct on a M6 32GB Mac?

Q: Can I run Llama-3.2-3B-Instruct on a M6 32GB Mac?

Yes — Llama-3.2-3B-Instruct fits in 32 GB of M6 unified memory at 8-bit with 32K context; recommended max ~90K tokens (keeps 20% headroom) · hard limit ~128K tokens. Computed from the model's official config.json by the open-source FitLLM engine.

✅ Yes — it fits at 32K context (8-bit) — 16.8 of 32 GB

Computed with the open FitLLM engine — accurate per-layer KV-cache modeling, not a naive estimate. Updated 2026-10-19.

Conditions: 8-bit · KV cache F16 · 32K context · macOS reserve (8.5 GB).

Memory breakdown (8-bit, F16 KV, 32K context)

Model weights	3.0 GB
KV cache	3.5 GB
Runtime + macOS	10.4 GB
Total used	16.8 / 32 GB
Free	15.2 GB

Context at 8-bit: recommended max ~90K tokens (keeps 20% headroom) · hard limit ~128K tokens. Unified memory is shared with macOS and your apps — the recommended max leaves 20% of it free.

Every quantization on M6 32GB

Quant	Weights	Recommended max ctx	Hard limit	Used @32K
4bit	1.5 GB	~101K	~128K	✅ 15.2 / 32 GB
8bit	3.0 GB	~90K	~128K	✅ 16.8 / 32 GB
16bit	6.0 GB	~68K	~110K	✅ 20.2 / 32 GB

Recommended max = the longest context that still keeps the 20% headroom macOS needs · hard limit = the longest that fits at all (KV cache F16).

Lower quants free memory at some output-quality cost — 4-bit is the common sweet spot for local use.

▶ Open the interactive calculator (this exact setup)

Embed this verdict

Live badge for your README or model card — recomputed by the engine at exactly these conditions, never stale:

[![fits: Llama-3.2-3B-Instruct on M6 32GB Mac](https://img.shields.io/endpoint?url=https%3A%2F%2Ffitllm.run%2Fapi%2Fbadge%3Fmodel%3DLlama-3.2-3B-Instruct%26ram%3D32%26quant%3D8%26ctx%3D32768%26kv%3D16)](https://fitllm.run/can-i-run/llama-3-2-3b-instruct-on-m6-32gb)

fit badge preview ← renders like this, live.

Or from your terminal (exit 0/1 — works as a pre-download guard):

npx fitllm "Llama-3.2-3B-Instruct" --mac 32 --quant 8 --ctx 32768 --kv 16

Why most calculators get this wrong

Llama-3.2-3B-Instruct's KV cache is computed per layer with its real head_dim and grouped-query head count — not the uniform "all layers × full context" shortcut most calculators use.

Other options

same Mac Models that fit in 32GB: Gemma 4 e2b, Gemma 4 e4b, Gemma 4 12b, Llama-3.2-3B-Instruct, Llama-3.1-8B-Instruct, MiniCPM5-1B, Qwen3-0.6B, Qwen3-1.7B, Llama-3.2-1B-Instruct, Gemma-3-1B-it, Spark-X2.5-4B, MiniCPM5-2B.

Reproduce it

Open math: fitllm-engine (MIT), from official config.json.

All numbers are computed by the open-source fitllm-engine (MIT) from official model config.json values — reproduce or audit them yourself. Estimates; real usage varies with runtime (llama.cpp / MLX / Ollama), driver and display. Found a mismatch? Report it. · FitLLM home