LLMCheck: Apple M3 Ultra (512GB), MLX, GLM 5.2 753B-A40B Q4_K_M decode
15.0 tok/s (decode, per-request)
Fields
id
benchmarks/llmcheck-m3-ultra-glm5-753b-q4km-mlx-decodeMetric
decode_tok_s
Value
15
Unit
tok/s (decode, per-request)
Model
GLM 5.2 753B-A40B
Methodology
THIRD-PARTY MEASURED (LLMCheck open dataset). M3 Ultra 512GB, MLX, GLM 5.2 753B-A40B Q4_K_M, TTFT 2.0 s, per-request.
Measured by
third_party
Reproducible
yes
notes
Q4_K_M row: 753B at 4-bit needs ~400GB for weights alone — only the 512GB SKU has the headroom, and decode is ~15 tok/s. The IQ1_S row (22 tok/s) shows the quant-vs-speed lever.
Notes
Q4_K_M row: 753B at 4-bit needs ~400GB for weights alone — only the 512GB SKU has the headroom, and decode is ~15 tok/s. The IQ1_S row (22 tok/s) shows the quant-vs-speed lever.