Best local LLM for summarization on a Apple M2 Pro (32GB) (2026)
All figures assume an f16 KV cache, a 0.6 GB display reserve on the GPU, and 64 GB of DDR5 system RAM for the offload tiers. Tune these in the calculator.
The verdict
Gemma 3 27B Q5_K_M
Fits on GPU at 32K · Summarization score 8/10 · ≈ 7 tok/s
Sliding-window attention keeps long context cheap, and the vision stack still beats most of 2026's newcomers.
llama-server -m gemma-3-27b-it-Q5_K_M.gguf -c 32768 --flash-attn -ngl 99
Worthy alternates
Qwen3.5 27B IQ4_XS
Fits on GPU · ≈ 9 tok/s · Summarization 8/10
The dense 24GB workhorse. If you want one model on a 3090 and no surprises, it's this.
Gemma 4 26B A4B IQ4_XS
Fits on GPU · ≈ 37 tok/s · Summarization 8/10
Google's fast MoE with native audio in. Nearly all of its weight sits in routed experts, so expert offload runs it comfortably on 12 GB cards.
Tune this for your exact RAM and settings in the calculator → · All models on the Apple M2 Pro (32GB)
Frequently asked questions
What is the best local LLM for summarization on a Apple M2 Pro (32GB)?
Gemma 3 27B at Q5_K_M — it scores 8/10 for summarization and runs as "Fits on GPU" at 32K context on the Apple M2 Pro (32GB).
How much context do I need for summarization?
We recommend 48K tokens for summarization (minimum 24K). These picks are computed at 32K.
How fast will it run on a Apple M2 Pro (32GB)?
Roughly 7 tokens/sec for Gemma 3 27B — usable for interactive use.
Do I need more than 24 GB of VRAM for summarization?
No — the pick above needs 23.1 GB of VRAM at 32K.
What settings should I use?
Start with our command: llama-server -m gemma-3-27b-it-Q5_K_M.gguf -c 32768 --flash-attn -ngl 99 — then tune context and KV quant in the fit calculator.