Cheapest box to run Llama-4-Behemoth at Q6_K
Meta’s Llama-4-Behemoth is 2000B parameters, 288B of them read per token. The number below is computed from that and each vendor’s published specs, not asserted. Switch the quant or the context and it recomputes in place.
Llama-4-Behemoth at Q6_K needs 1643GB.
Nothing on the Grid runs Llama-4-Behemoth at Q6_K today. The largest Config, Mac Studio, M5 Ultra, 512GB, is 1139GB short.
Announced, weights not released. Parameter counts are Meta's; the K/V figure is an estimate from the Llama 4 family and is marked as one.
- Mac mini, M6, 16GB$1,099Won't fit−1630GB headroom
- Mac mini, M6, 24GB$1,299Won't fit−1622GB headroom
- Mac mini, M6, 32GB$1,499Won't fit−1614GB headroom
- Mac mini, M5 Pro, 24GB$1,699Won't fit−1622GB headroom
- Mac mini, M5 Pro, 48GB$2,499Won't fit−1599GB headroom
- Mac Studio, M5 Max, 36GB$2,499Won't fit−1610GB headroom
- Mac mini, M5 Pro, 64GB$2,899Won't fit−1584GB headroom
- Mac Studio, M5 Max, 48GB$3,099Won't fit−1599GB headroom
- Mac Studio, M5 Max, 64GB$3,499Won't fit−1584GB headroom
- Won't fit−1523GB headroom
- Won't fit−1523GB headroom
- Won't fit−1555GB headroom
- GeForce RTX 5090, 32GB$5,699.99+hostWon't fit−1612GB headroom
- Won't fit−1395GB headroom
- Won't fit−1139GB headroom
| Config | Memory | Price | $/GB | Fit | Headroom | Est. speed |
|---|---|---|---|---|---|---|
| Mac mini, M6 | 16GB | $1,099 | $69 | Won't fit | −1630GB | — |
| Mac mini, M6 | 24GB | $1,299 | $54 | Won't fit | −1622GB | — |
| Mac mini, M6 | 32GB | $1,499 | $47 | Won't fit | −1614GB | — |
| Mac mini, M5 Pro | 24GB | $1,699 | $71 | Won't fit | −1622GB | — |
| Mac mini, M5 Pro | 48GB | $2,499 | $52 | Won't fit | −1599GB | — |
| Mac Studio, M5 Max | 36GB | $2,499 | $69 | Won't fit | −1610GB | — |
| Mac mini, M5 Pro | 64GB | $2,899 | $45 | Won't fit | −1584GB | — |
| Mac Studio, M5 Max | 48GB | $3,099 | $65 | Won't fit | −1599GB | — |
| Mac Studio, M5 Max | 64GB | $3,499 | $55 | Won't fit | −1584GB | — |
| DGX Spark, GB10 Grace Blackwell | 128GB | $4,699 | $37 | Won't fit | −1523GB | — |
| Mac Studio, M5 Max | 128GB | $5,099 | $40 | Won't fit | −1523GB | — |
| Mac Studio, M5 Ultra | 96GB | $5,499 | $57 | Won't fit | −1555GB | — |
| GeForce RTX 5090 | 32GB | $5,699.99+host | $178 | Won't fit | −1612GB | — |
| Mac Studio, M5 Ultra | 256GB | $9,499 | $37 | Won't fit | −1395GB | — |
| Mac Studio, M5 Ultra | 512GB | — | — | Won't fit | −1139GB | — |
GeForce RTX 5090 is a card, not a computer: its price is the US street median plus $1,000 for the minimum host machine. NVIDIA's $1,999 MSRP is not a price anyone pays.
memory = 2000B × 0.820 B/param (6.56 bits) + 0.250 GB/1K × 8K K/V + 1.2 GB runtime
speed = bandwidth × 0.5 ÷ (288B active × 0.820 B/param) = 236.2 GB read per token
usable = memory − operating system reserve (8% on a Mac, 6% on a DGX Spark, 4% on a card)
Bandwidth is the vendor’s published figure. Every speed is an estimate and says so. Methodology.