LLMCheck: Apple M5 Max, MLX, GLM-4.5-Air 106B-A?B Q4_K_M decode
34.0 tok/s (decode, per-request)
Fields
id
benchmarks/llmcheck-m5-max-glm45-air-mlx-decodeMetric
decode_tok_s
Value
34
Unit
tok/s (decode, per-request)
Model
GLM-4.5-Air
Methodology
THIRD-PARTY MEASURED (LLMCheck open dataset). M5 Max, MLX, GLM-4.5-Air (106B MoE) Q4_K_M, TTFT 0.9 s, per-request.
Measured by
third_party
Reproducible
yes
notes
Second large-MoE data point; the repo's GLM-4.6-Air target model is in the same class.
Notes
Second large-MoE data point; the repo's GLM-4.6-Air target model is in the same class.