M2 Max 64 64GB

// weights Q3_K_M for Command R+ (104B params)
weights = params × bits ÷ 8
        = 104 × 3.44 ÷ 8
        = 44.72 GB

// budget on M2 Max 64 (64GB) at ctx 8K, conc 1, 15% safety
kv_cache  = 0.05 GB    (1× at ctx 8K)
overhead  = 2.50 GB    (runtime, cuda, allocator)
safety    = 9.60 GB    (15% of 64GB)
budget    = vram − safety − kv − overhead
          = 64 − 9.60 − 0.05 − 2.50
          = 51.85 GB

// fit decision
44.72 ≤ 51.85  → FITS
headroom  = 7.13 GB of weights budget left

// weights Q5_K_M for Qwen 2.5 72B (72.7B params)
weights = params × bits ÷ 8
        = 72.7 × 5.5 ÷ 8
        = 49.98 GB

// budget on M2 Max 64 (64GB) at ctx 8K, conc 1, 15% safety
kv_cache  = 0.05 GB    (1× at ctx 8K)
overhead  = 2.50 GB    (runtime, cuda, allocator)
safety    = 9.60 GB    (15% of 64GB)
budget    = vram − safety − kv − overhead
          = 64 − 9.60 − 0.05 − 2.50
          = 51.85 GB

// fit decision
49.98 ≤ 51.85  → FITS
headroom  = 1.87 GB of weights budget left

// weights Q5_K_M for Llama 3.3 70B (70.6B params)
weights = params × bits ÷ 8
        = 70.6 × 5.5 ÷ 8
        = 48.54 GB

// budget on M2 Max 64 (64GB) at ctx 8K, conc 1, 15% safety
kv_cache  = 0.05 GB    (1× at ctx 8K)
overhead  = 2.50 GB    (runtime, cuda, allocator)
safety    = 9.60 GB    (15% of 64GB)
budget    = vram − safety − kv − overhead
          = 64 − 9.60 − 0.05 − 2.50
          = 51.85 GB

// fit decision
48.54 ≤ 51.85  → FITS
headroom  = 3.31 GB of weights budget left

// weights Q8_0 for Mixtral 8x7B (46.7B params)
weights = params × bits ÷ 8
        = 46.7 × 8.5 ÷ 8
        = 49.62 GB

// budget on M2 Max 64 (64GB) at ctx 8K, conc 1, 15% safety
kv_cache  = 0.05 GB    (1× at ctx 8K)
overhead  = 2.50 GB    (runtime, cuda, allocator)
safety    = 9.60 GB    (15% of 64GB)
budget    = vram − safety − kv − overhead
          = 64 − 9.60 − 0.05 − 2.50
          = 51.85 GB

// fit decision
49.62 ≤ 51.85  → FITS
headroom  = 2.23 GB of weights budget left

// weights Q8_0 for Qwen 3.6 35B A3B (35B params)
weights = params × bits ÷ 8
        = 35 × 8.5 ÷ 8
        = 37.19 GB

// budget on M2 Max 64 (64GB) at ctx 8K, conc 1, 15% safety
kv_cache  = 0.05 GB    (1× at ctx 8K)
overhead  = 2.50 GB    (runtime, cuda, allocator)
safety    = 9.60 GB    (15% of 64GB)
budget    = vram − safety − kv − overhead
          = 64 − 9.60 − 0.05 − 2.50
          = 51.85 GB

// fit decision
37.19 ≤ 51.85  → FITS
headroom  = 14.66 GB of weights budget left

// weights Q8_0 for Yi 34B (34B params)
weights = params × bits ÷ 8
        = 34 × 8.5 ÷ 8
        = 36.13 GB

// budget on M2 Max 64 (64GB) at ctx 8K, conc 1, 15% safety
kv_cache  = 0.05 GB    (1× at ctx 8K)
overhead  = 2.50 GB    (runtime, cuda, allocator)
safety    = 9.60 GB    (15% of 64GB)
budget    = vram − safety − kv − overhead
          = 64 − 9.60 − 0.05 − 2.50
          = 51.85 GB

// fit decision
36.13 ≤ 51.85  → FITS
headroom  = 15.72 GB of weights budget left

// weights Q8_0 for Qwen 2.5 32B (32.5B params)
weights = params × bits ÷ 8
        = 32.5 × 8.5 ÷ 8
        = 34.53 GB

// budget on M2 Max 64 (64GB) at ctx 8K, conc 1, 15% safety
kv_cache  = 0.05 GB    (1× at ctx 8K)
overhead  = 2.50 GB    (runtime, cuda, allocator)
safety    = 9.60 GB    (15% of 64GB)
budget    = vram − safety − kv − overhead
          = 64 − 9.60 − 0.05 − 2.50
          = 51.85 GB

// fit decision
34.53 ≤ 51.85  → FITS
headroom  = 17.32 GB of weights budget left

// weights Q8_0 for Qwen 2.5 Coder 32B (32.5B params)
weights = params × bits ÷ 8
        = 32.5 × 8.5 ÷ 8
        = 34.53 GB

// budget on M2 Max 64 (64GB) at ctx 8K, conc 1, 15% safety
kv_cache  = 0.05 GB    (1× at ctx 8K)
overhead  = 2.50 GB    (runtime, cuda, allocator)
safety    = 9.60 GB    (15% of 64GB)
budget    = vram − safety − kv − overhead
          = 64 − 9.60 − 0.05 − 2.50
          = 51.85 GB

// fit decision
34.53 ≤ 51.85  → FITS
headroom  = 17.32 GB of weights budget left

// weights Q8_0 for Qwen3 30B A3B (30.5B params)
weights = params × bits ÷ 8
        = 30.5 × 8.5 ÷ 8
        = 32.41 GB

// budget on M2 Max 64 (64GB) at ctx 8K, conc 1, 15% safety
kv_cache  = 0.05 GB    (1× at ctx 8K)
overhead  = 2.50 GB    (runtime, cuda, allocator)
safety    = 9.60 GB    (15% of 64GB)
budget    = vram − safety − kv − overhead
          = 64 − 9.60 − 0.05 − 2.50
          = 51.85 GB

// fit decision
32.41 ≤ 51.85  → FITS
headroom  = 19.44 GB of weights budget left

// weights Q8_0 for Qwen 3.6 27B (27B params)
weights = params × bits ÷ 8
        = 27 × 8.5 ÷ 8
        = 28.69 GB

// budget on M2 Max 64 (64GB) at ctx 8K, conc 1, 15% safety
kv_cache  = 0.05 GB    (1× at ctx 8K)
overhead  = 2.50 GB    (runtime, cuda, allocator)
safety    = 9.60 GB    (15% of 64GB)
budget    = vram − safety − kv − overhead
          = 64 − 9.60 − 0.05 − 2.50
          = 51.85 GB

// fit decision
28.69 ≤ 51.85  → FITS
headroom  = 23.16 GB of weights budget left

// weights Q8_0 for Gemma 4 26B A4B (26B params)
weights = params × bits ÷ 8
        = 26 × 8.5 ÷ 8
        = 27.63 GB

// budget on M2 Max 64 (64GB) at ctx 8K, conc 1, 15% safety
kv_cache  = 0.05 GB    (1× at ctx 8K)
overhead  = 2.50 GB    (runtime, cuda, allocator)
safety    = 9.60 GB    (15% of 64GB)
budget    = vram − safety − kv − overhead
          = 64 − 9.60 − 0.05 − 2.50
          = 51.85 GB

// fit decision
27.63 ≤ 51.85  → FITS
headroom  = 24.22 GB of weights budget left

// weights FP16/BF16 for Mistral Small 3 (24B params)
weights = params × bits ÷ 8
        = 24 × 16 ÷ 8
        = 48.00 GB

// budget on M2 Max 64 (64GB) at ctx 8K, conc 1, 15% safety
kv_cache  = 0.05 GB    (1× at ctx 8K)
overhead  = 2.50 GB    (runtime, cuda, allocator)
safety    = 9.60 GB    (15% of 64GB)
budget    = vram − safety − kv − overhead
          = 64 − 9.60 − 0.05 − 2.50
          = 51.85 GB

// fit decision
48.00 ≤ 51.85  → FITS
headroom  = 3.85 GB of weights budget left