Cheapest box to run Llama-4-Scout-17B-16E at Q8_0
Meta’s Llama-4-Scout-17B-16E is 109B parameters, 17B 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-Scout-17B-16E at Q8_0 needs 119GB.
Cheapest Config that fits: Mac Studio, M5 Ultra, 256GB. Fits, 129GB headroom. Est. 33.2 tok/s.
Best offer today: $9,499 at Apple — in stock, checked 10 hours ago.
Collecting price history — 2 of 90 days. No timing verdict yet.
Offers — Mac Studio, M5 Ultra, 256GB, 1TB
- Mac mini, M6, 16GB$1,099Won't fit−106GB headroom
- Mac mini, M6, 24GB$1,299Won't fit−98GB headroom
- Mac mini, M6, 32GB$1,499Won't fit−90GB headroom
- Mac mini, M5 Pro, 24GB$1,699Won't fit−98GB headroom
- Mac mini, M5 Pro, 48GB$2,499Won't fit−74GB headroom
- Mac Studio, M5 Max, 36GB$2,499Won't fit−86GB headroom
- Mac mini, M5 Pro, 64GB$2,899Won't fit−60GB headroom
- GeForce RTX 5090, 32GB$2,999.99+hostWon't fit−88GB headroom
- Mac Studio, M5 Max, 48GB$3,099Won't fit−74GB headroom
- Mac Studio, M5 Max, 64GB$3,499Won't fit−60GB headroom
- Tight2GB headroomest. 7.6 tok/s
- Tight2GB headroomest. 7.6 tok/s
- Tight1GB headroomest. 17.0 tok/s
- Won't fit−30GB headroom
- Fits129GB headroomest. 33.2 tok/s
- Fits385GB headroomest. 33.2 tok/s
| Config | Memory | Price | $/GB | Fit | Headroom | Est. speed |
|---|---|---|---|---|---|---|
| Mac mini, M6 | 16GB | $1,099 | $69 | Won't fit | −106GB | — |
| Mac mini, M6 | 24GB | $1,299 | $54 | Won't fit | −98GB | — |
| Mac mini, M6 | 32GB | $1,499 | $47 | Won't fit | −90GB | — |
| Mac mini, M5 Pro | 24GB | $1,699 | $71 | Won't fit | −98GB | — |
| Mac mini, M5 Pro | 48GB | $2,499 | $52 | Won't fit | −74GB | — |
| Mac Studio, M5 Max | 36GB | $2,499 | $69 | Won't fit | −86GB | — |
| Mac mini, M5 Pro | 64GB | $2,899 | $45 | Won't fit | −60GB | — |
| GeForce RTX 5090 | 32GB | $2,999.99+host | $94 | Won't fit | −88GB | — |
| Mac Studio, M5 Max | 48GB | $3,099 | $65 | Won't fit | −74GB | — |
| Mac Studio, M5 Max | 64GB | $3,499 | $55 | Won't fit | −60GB | — |
| Ascent GX10, GB10 Grace Blackwell | 128GB | $3,999.99 | $31 | Tight | 2GB | 7.6 tok/s |
| DGX Spark, GB10 Grace Blackwell | 128GB | $4,699 | $37 | Tight | 2GB | 7.6 tok/s |
| Mac Studio, M5 Max | 128GB | $5,099 | $40 | Tight | 1GB | 17.0 tok/s |
| Mac Studio, M5 Ultra | 96GB | $5,499 | $57 | Won't fit | −30GB | — |
| Mac Studio, M5 Ultra | 256GB | $9,499 | $37 | Fits | 129GB | 33.2 tok/s |
| Mac Studio, M5 Ultra | 512GB | — | — | Fits | 385GB | 33.2 tok/s |
GeForce RTX 5090 is a card, not a computer: every price shown for it is the cheapest listing we found plus $1,000 for the minimum host machine.
memory = 109B × 1.063 B/param (8.5 bits) + 0.197 GB/1K × 8K K/V + 1.2 GB runtime
speed = bandwidth × 0.5 ÷ (17B active × 1.063 B/param) = 18.1 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.
What else Mac Studio M5 Ultra 256GB runs
- Llama-4-Maverick-17B-128EQ4_K_M, est. 58.5 tok/s
- Qwen3-235B-A22BQ6_K, est. 33.3 tok/s
- Mistral-Large-2411Q8_0, est. 6.0 tok/s
- gpt-oss-120bBF16, est. 58.8 tok/s
- Llama-4-Scout-17B-16EBF16, est. 17.6 tok/s
- Llama-3.3-70BBF16, est. 5.6 tok/s
Best quant that still fits at 8K context, largest model first. Full list on the Config page.
Bandwidth 1.2TB/s, usable memory 248GB of 256GB.