M5 Max 128 128GB

// weights Q5_K_M for Mixtral 8x22B (141B params)
weights = params × bits ÷ 8
        = 141 × 5.5 ÷ 8
        = 96.94 GB

// budget on M5 Max 128 (128GB) at ctx 8K, conc 1, 15% safety
kv_cache  = 0.05 GB    (1× at ctx 8K)
overhead  = 2.50 GB    (runtime, cuda, allocator)
safety    = 19.20 GB    (15% of 128GB)
budget    = vram − safety − kv − overhead
          = 128 − 19.20 − 0.05 − 2.50
          = 106.25 GB

// fit decision
96.94 ≤ 106.25  → FITS
headroom  = 9.31 GB of weights budget left

// weights Q6_K for gpt-oss 120B (117B params)
weights = params × bits ÷ 8
        = 117 × 6.56 ÷ 8
        = 95.94 GB

// budget on M5 Max 128 (128GB) at ctx 8K, conc 1, 15% safety
kv_cache  = 0.05 GB    (1× at ctx 8K)
overhead  = 2.50 GB    (runtime, cuda, allocator)
safety    = 19.20 GB    (15% of 128GB)
budget    = vram − safety − kv − overhead
          = 128 − 19.20 − 0.05 − 2.50
          = 106.25 GB

// fit decision
95.94 ≤ 106.25  → FITS
headroom  = 10.31 GB of weights budget left

// weights Q6_K for Command R+ (104B params)
weights = params × bits ÷ 8
        = 104 × 6.56 ÷ 8
        = 85.28 GB

// budget on M5 Max 128 (128GB) at ctx 8K, conc 1, 15% safety
kv_cache  = 0.05 GB    (1× at ctx 8K)
overhead  = 2.50 GB    (runtime, cuda, allocator)
safety    = 19.20 GB    (15% of 128GB)
budget    = vram − safety − kv − overhead
          = 128 − 19.20 − 0.05 − 2.50
          = 106.25 GB

// fit decision
85.28 ≤ 106.25  → FITS
headroom  = 20.97 GB of weights budget left

// weights Q8_0 for Qwen 2.5 72B (72.7B params)
weights = params × bits ÷ 8
        = 72.7 × 8.5 ÷ 8
        = 77.24 GB

// budget on M5 Max 128 (128GB) at ctx 8K, conc 1, 15% safety
kv_cache  = 0.05 GB    (1× at ctx 8K)
overhead  = 2.50 GB    (runtime, cuda, allocator)
safety    = 19.20 GB    (15% of 128GB)
budget    = vram − safety − kv − overhead
          = 128 − 19.20 − 0.05 − 2.50
          = 106.25 GB

// fit decision
77.24 ≤ 106.25  → FITS
headroom  = 29.00 GB of weights budget left

// weights Q8_0 for Llama 3.3 70B (70.6B params)
weights = params × bits ÷ 8
        = 70.6 × 8.5 ÷ 8
        = 75.01 GB

// budget on M5 Max 128 (128GB) at ctx 8K, conc 1, 15% safety
kv_cache  = 0.05 GB    (1× at ctx 8K)
overhead  = 2.50 GB    (runtime, cuda, allocator)
safety    = 19.20 GB    (15% of 128GB)
budget    = vram − safety − kv − overhead
          = 128 − 19.20 − 0.05 − 2.50
          = 106.25 GB

// fit decision
75.01 ≤ 106.25  → FITS
headroom  = 31.24 GB of weights budget left

// weights FP16/BF16 for Mixtral 8x7B (46.7B params)
weights = params × bits ÷ 8
        = 46.7 × 16 ÷ 8
        = 93.40 GB

// budget on M5 Max 128 (128GB) at ctx 8K, conc 1, 15% safety
kv_cache  = 0.05 GB    (1× at ctx 8K)
overhead  = 2.50 GB    (runtime, cuda, allocator)
safety    = 19.20 GB    (15% of 128GB)
budget    = vram − safety − kv − overhead
          = 128 − 19.20 − 0.05 − 2.50
          = 106.25 GB

// fit decision
93.40 ≤ 106.25  → FITS
headroom  = 12.85 GB of weights budget left

// weights FP16/BF16 for Qwen 3.6 35B A3B (35B params)
weights = params × bits ÷ 8
        = 35 × 16 ÷ 8
        = 70.00 GB

// budget on M5 Max 128 (128GB) at ctx 8K, conc 1, 15% safety
kv_cache  = 0.05 GB    (1× at ctx 8K)
overhead  = 2.50 GB    (runtime, cuda, allocator)
safety    = 19.20 GB    (15% of 128GB)
budget    = vram − safety − kv − overhead
          = 128 − 19.20 − 0.05 − 2.50
          = 106.25 GB

// fit decision
70.00 ≤ 106.25  → FITS
headroom  = 36.25 GB of weights budget left

// weights FP16/BF16 for Yi 34B (34B params)
weights = params × bits ÷ 8
        = 34 × 16 ÷ 8
        = 68.00 GB

// budget on M5 Max 128 (128GB) at ctx 8K, conc 1, 15% safety
kv_cache  = 0.05 GB    (1× at ctx 8K)
overhead  = 2.50 GB    (runtime, cuda, allocator)
safety    = 19.20 GB    (15% of 128GB)
budget    = vram − safety − kv − overhead
          = 128 − 19.20 − 0.05 − 2.50
          = 106.25 GB

// fit decision
68.00 ≤ 106.25  → FITS
headroom  = 38.25 GB of weights budget left

// weights FP16/BF16 for Qwen 2.5 32B (32.5B params)
weights = params × bits ÷ 8
        = 32.5 × 16 ÷ 8
        = 65.00 GB

// budget on M5 Max 128 (128GB) at ctx 8K, conc 1, 15% safety
kv_cache  = 0.05 GB    (1× at ctx 8K)
overhead  = 2.50 GB    (runtime, cuda, allocator)
safety    = 19.20 GB    (15% of 128GB)
budget    = vram − safety − kv − overhead
          = 128 − 19.20 − 0.05 − 2.50
          = 106.25 GB

// fit decision
65.00 ≤ 106.25  → FITS
headroom  = 41.25 GB of weights budget left

// weights FP16/BF16 for Qwen 2.5 Coder 32B (32.5B params)
weights = params × bits ÷ 8
        = 32.5 × 16 ÷ 8
        = 65.00 GB

// budget on M5 Max 128 (128GB) at ctx 8K, conc 1, 15% safety
kv_cache  = 0.05 GB    (1× at ctx 8K)
overhead  = 2.50 GB    (runtime, cuda, allocator)
safety    = 19.20 GB    (15% of 128GB)
budget    = vram − safety − kv − overhead
          = 128 − 19.20 − 0.05 − 2.50
          = 106.25 GB

// fit decision
65.00 ≤ 106.25  → FITS
headroom  = 41.25 GB of weights budget left

// weights FP16/BF16 for Qwen3 30B A3B (30.5B params)
weights = params × bits ÷ 8
        = 30.5 × 16 ÷ 8
        = 61.00 GB

// budget on M5 Max 128 (128GB) at ctx 8K, conc 1, 15% safety
kv_cache  = 0.05 GB    (1× at ctx 8K)
overhead  = 2.50 GB    (runtime, cuda, allocator)
safety    = 19.20 GB    (15% of 128GB)
budget    = vram − safety − kv − overhead
          = 128 − 19.20 − 0.05 − 2.50
          = 106.25 GB

// fit decision
61.00 ≤ 106.25  → FITS
headroom  = 45.25 GB of weights budget left

// weights FP16/BF16 for Qwen 3.6 27B (27B params)
weights = params × bits ÷ 8
        = 27 × 16 ÷ 8
        = 54.00 GB

// budget on M5 Max 128 (128GB) at ctx 8K, conc 1, 15% safety
kv_cache  = 0.05 GB    (1× at ctx 8K)
overhead  = 2.50 GB    (runtime, cuda, allocator)
safety    = 19.20 GB    (15% of 128GB)
budget    = vram − safety − kv − overhead
          = 128 − 19.20 − 0.05 − 2.50
          = 106.25 GB

// fit decision
54.00 ≤ 106.25  → FITS
headroom  = 52.25 GB of weights budget left