Benchmarks · community-measured

qwen3-8b-4bit on Apple M4 Pro

Real measured local-inference performance for qwen3-8b-4bit on an Apple M4 Pro (48 GB unified memory), from the rapid-mlx community benchmark corpus. Headline submission 2026-08-27, measured on rapid-mlx 0.13.1.

48.3
tok/s decode
446
tok/s prefill
4.8 s
first token
8.0 GB
peak memory

A single community run so far — treat it as one data point, not a distribution. Reproduce it below and submit yours to firm it up.

All measurements for this pair

RAMAccelDecodePrefillTTFTPeak memRunsrapid-mlxLatest
48 GB48.3 tok/s446 tok/s4.8 s8.0 GB10.13.12026-08-27

Standardized single-request benchmark: 5 measured rounds after 1 discarded warmup, greedy sampling, prefix cache disabled. Raw corpus: /api/benchmarks.

Reproduce this number

Install rapid-mlx, run the same suite on your Mac, and submit the result — it lands on the community leaderboard and updates this page.

curl -fsSL https://rapidmlx.com/install.sh | bash
rapid-mlx bench qwen3-8b-4bit --tier speed --submit

More

qwen3-8b-4bit model page · Qwen family · Interactive leaderboard · Alias catalog