Best local LLMs for an M1 Max Mac with 64 GB

7 models have been measured on an Apple M1 Max with 64 GB of unified memory. Every number below was measured on that chip and memory size, by the Rapid-MLX community or by us. None is scaled from another Mac.

What Rapid-MLX picks for 64 GB

These two come from the engine's recommendation policy for its 64 GB tier (rapid/default/text-generation/ram-v1): the same picks rapid-mlx recipe, the installer and the desktop app show. They are the policy's choice, not a ranking of the measurements below. How the tiers are chosen.

Measured on this Mac

Decode speed is how fast the model writes once it has read your prompt; first-token time is how long it reads before it starts. Each section has its own scale and method, so compare rows within a section.

Community runs

Current community benchmark: a 512-token prompt, then 128 generated tokens. Median across runs; each run on the contributor's own Mac.

• bonsai-1.7b-2bit217tok/s
  first token 223 ms · peak memory 2.3 GB · rapid-mlx 0.15.5 · 1 run · 2026-10-04
• gpt-oss-20b69.0tok/s
  first token 804 ms · peak memory 12.4 GB · rapid-mlx 0.15.5 · 1 run · 2026-10-04
• gemma-4-e4b-4bit67.7tok/s
  first token 480 ms · peak memory 5.2 GB · rapid-mlx 0.15.5 · 1 run · 2026-10-04
• qwen3.6-35b62.5tok/s
  first token 984 ms · peak memory 20.3 GB · rapid-mlx 0.15.5 · 1 run · 2026-10-04
• qwen3.5-9b-4bit45.7tok/s
  first token 1.6 s · peak memory 6.7 GB · rapid-mlx 0.15.5 · 1 run · 2026-10-03
• qwen3.8-27b-4bit12.1tok/s
  first token 7.1 s · peak memory 18.6 GB · rapid-mlx 0.13.4, 0.14.3, 0.15.2, 0.15.5 · 5 runs · 2026-10-04

Older community benchmark

The previous community suite (short prompt). Its protocol differs from the current one, so compare these rows with each other, not with the sections above.

• qwen3.5-4b-4bit91.9tok/s
  first token 761 ms · rapid-mlx 0.13.2 · 1 run · 2026-09-01 · details

Run one

rapid-mlx recipe detects your Mac's memory and prints the pick with its command. Any model above runs with rapid-mlx serve <alias> and serves an OpenAI- and Anthropic-compatible API on http://localhost:8000.

$ curl -fsSL https://rapidmlx.com/install.sh | bash
$ rapid-mlx recipe
$ rapid-mlx serve qwen3.8-27b-4bit

Add your Mac's numbers

Have an M1 Max Mac with 64 GB? Run the benchmark and share it. It lands on the live leaderboard view for this Mac and on this page at its next build.

$ rapid-mlx benchmark run qwen3.8-27b-4bit
$ rapid-mlx benchmark share <run-id>

More

• Every Mac with enough measurements
• The best local LLM for your Mac, by memory
• Hardware tiers: the engine's picks for every memory size
• Rapid-MLX vs mlx-lm, measured task by task
• Use these models with a coding agent