Silicon Score
← All benchmarks

Silicon Score › Chips › M4 Max (128 GB)

M4 Max (128 GB) — LLM Benchmarks

Measured LLM inference benchmarks for M4 Max (128 GB). Tokens per second across 8 models and multiple quantizations. Real runs, not estimates.

8Benchmark rows
8Models tested
184.4Fastest avg tok/s (Qwen 3 0.6B)
0Factory-lab verified rows

Benchmark results for M4 Max (128 GB)

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

ModelQuantRAM req.ContextAvg tok/sPrompt tok/sRuntimeSource
Qwen 3 0.6BQ8_0—10k184.4 tok/s—LM Studioref
Gemma 3 4BQ4_0—4k100.5 tok/s—LM Studioref
Qwen 3 30B A3BQ4_K_M—10k70.2 tok/s—LM Studioref
Qwen 3 8BQ4_K_M—10k63.1 tok/s—LM Studioref
Qwen 2.5 7B InstructQ8_0—10k49.7 tok/s—LM Studioref
Gemma 3 27BQ8_0—131k14.5 tok/s—LM Studioref
Qwen 3 32BQ4_K_M—10k11.7 tok/s—LM Studioref
Qwen 3 235B A22BQ4_K_M—10k8.1 tok/s—LM Studioref

Models tested on this chip

Qwen 3 0.6BGemma 3 4BQwen 3 30B A3BQwen 3 8BQwen 2.5 7B InstructGemma 3 27BQwen 3 32BQwen 3 235B A22B

Data

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

Data sourced from factory lab measurements and community reference runs. See all chips →