MLX-LM benchmark data for Qwen3-4B and Qwen3-30B-A3B across precisions (mlx-lm PR #552)

Fields

id

sources/bench-apple-mlx-lm-qwen3-benchmarks

URL

https://github.com/ml-explore/mlx-lm/pull/552

Publisher

ml-explore (Apple) / community

Kind

benchmark

Accessed

2026-10-03

notes

GitHub PR #552 in ml-explore/mlx-lm adding LLM benchmarks. Published 2025-10-14 by awni. Adds benchmark info to the repo. Verified 2026-10-10: Read PR 'LLM Benchmarks' (https://github.com/ml-explore/mlx-lm/pull/552); confirms PR by awni merged BENCHMARKS.md with benchmark instructions and model benchmark tables.

Notes

GitHub PR #552 in ml-explore/mlx-lm adding LLM benchmarks. Published 2025-10-14 by awni. Adds benchmark info to the repo. Verified 2026-10-10: Read PR 'LLM Benchmarks' (https://github.com/ml-explore/mlx-lm/pull/552); confirms PR by awni merged BENCHMARKS.md with benchmark instructions and model benchmark tables.

Cross-references

This record references no other records.