Silicon Score Apple Silicon local LLM rankings
← Canonical rankings

Silicon Score › Chips › M5 Max (64 GB)

M5 Max (64 GB) — benchmark record

M5 Max (64 GB) local LLM benchmarks on Apple Silicon Mac with 23 published rows across 20 models. Peak published speed is 148.0 tok/s on Qwen3.5-4B. Published runtimes include MLX, Ollama.

23Benchmark rows
20Models tested
148.0Fastest avg tok/s (Qwen3.5-4B)
0Lab benchmarks

Quick take

Best published speed here is 148.0 tok/s on Qwen3.5-4B at Q4_K - Medium. This page is evidence, not the full buying answer.

Based on 23 external benchmarks; no lab runs yet.

Published runtimes: MLX, Ollama.

Need a recommendation? Use BuyChecking a target model? Use FitWant a setup answer? Use RunBrowse Macs by exact hardwareNeed the full audit trail? Use Bench

Current published coverage

Published model coverage includes Qwen3.5-4B, Phi-4 Mini Instruct 3.8B, Qwen 3 4B, Gemma 3 4B plus 16 more published models. Published runtime coverage on this chip includes MLX, Ollama. Fastest published row is 148.0 tok/s on Qwen3.5-4B at Q4_K - Medium.

Qwen3.5-4BPhi-4 Mini Instruct 3.8BQwen 3 4BGemma 3 4BMistral 7B v0.3Qwen3.5-9B

Other published M5 Max variants: 48 GB · 128 GB

Macs shipping with the M5 Max (64 GB)

MacBook Pro M5 Max 128GB 16-inch

This chip is part of a family. Compare all M5 Max RAM variants →

Raw benchmark rows for M5 Max (64 GB)

Rows sorted by avg tok/s descending. Click source badge to see original measurement page.

ModelQuantRAM req.ContextAvg tok/sPrompt tok/sRuntimeSource
Qwen3.5-4BQ4_K - Medium——148.0 tok/s—MLXref
Phi-4 Mini Instruct 3.8BQ4_K - Medium——142.0 tok/s—Ollamaref
Qwen 3 4BQ4_K - Medium——135.0 tok/s—Ollamaref
Gemma 3 4BQ4_K - Medium——132.0 tok/s—Ollamaref
Mistral 7B v0.3Q4_K - Medium——122.0 tok/s—Ollamaref
Phi-4 Mini Instruct 3.8BQ8_0——112.0 tok/s—MLXref
Qwen3.5-9BQ4_K - Medium——105.0 tok/s—Ollamaref
Ministral 3 8BQ4_K - Medium——98.0 tok/s—Ollamaref
DeepSeek R1 Distill Llama 8BQ4_K - Medium——97.0 tok/s—Ollamaref
Mistral 7B v0.3Q8_0——88.0 tok/s—MLXref
Llama 3.1 8BQ8_0——82.0 tok/s—Ollamaref
DeepSeek R1 Distill Llama 8BQ8_0——75.0 tok/s—MLXref
Gemma 3 12BQ4_K - Medium——68.0 tok/s—Ollamaref
Phi-4 14BQ4_K - Medium——62.0 tok/s—MLXref
Qwen 3 30B-A3BQ4_K - Medium——62.0 tok/s—Ollamaref
Qwen 3 14BQ4_K - Medium——58.0 tok/s—Ollamaref
Ministral 3 14BQ4_K - Medium——58.0 tok/s—Ollamaref
Qwen3.5-35B-A3BQ4_K - Medium——52.0 tok/s—MLXref
Qwen3.6-35B-A3BQ4_K - Medium——48.0 tok/s—Ollamaref
Gemma 3 27BQ4_K - Medium——42.0 tok/s—Ollamaref
Nemotron Cascade 2 30B-A3BQ4_K - Medium——35.0 tok/s—Ollamaref
DeepSeek R1 Distill Qwen 32BQ4_K - Medium——27.0 tok/s—Ollamaref
Gemma 4 31BQ4_K - Medium——22.0 tok/s—Ollamaref

Models tested on this chip

Qwen3.5-4BPhi-4 Mini Instruct 3.8BQwen 3 4BGemma 3 4BMistral 7B v0.3Qwen3.5-9BMinistral 3 8BDeepSeek R1 Distill Llama 8BLlama 3.1 8BGemma 3 12BPhi-4 14BQwen 3 30B-A3BQwen 3 14BMinistral 3 14BQwen3.5-35B-A3BQwen3.6-35B-A3BGemma 3 27BNemotron Cascade 2 30B-A3BDeepSeek R1 Distill Qwen 32BGemma 4 31B

Next step

Use Rankings when you need the best overall answer for a Mac, and keep this page open when you need the evidence behind that recommendation.

If you are comparing local Apple Silicon against rented cloud hardware, use AI Data Center Index for current GPU rental context.

Data

benchmarks.json — full dataset  ·  chips.json — chip summaries  ·  benchmarks.csv — CSV export

Data from in-house lab measurements plus community-published benchmarks. See all chip families →