Cheapest box to run Devstral-Small-2507 at BF16
Mistral’s Devstral-Small-2507 is 24B parameters, dense. 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.
Devstral-Small-2507 at BF16 needs 51GB.
Cheapest Config that fits: Mac mini, M5 Pro, 64GB. Fits, 8GB headroom. Est. 4.2 tok/s.
Best offer today: $2,899 at Apple — in stock, checked 4 minutes ago.
Collecting price history — 3 of 90 days. No timing verdict yet.
Offers — Mac mini, M5 Pro, 64GB, 512GB
- Mac mini, M6, 16GB$1,099Won't fit−38GB headroom
- Mac mini, M6, 24GB$1,299Won't fit−30GB headroom
- Mac mini, M6, 32GB$1,499Won't fit−22GB headroom
- Mac mini, M5 Pro, 24GB$1,699Won't fit−30GB headroom
- Mac mini, M5 Pro, 48GB$2,499Won't fit−6GB headroom
- Mac Studio, M5 Max, 36GB$2,499Won't fit−18GB headroom
- Mac mini, M5 Pro, 64GB$2,899Fits8GB headroomest. 4.2 tok/s
- GeForce RTX 5090, 32GB$2,999.99+hostWon't fit−20GB headroom
- Fits8GB headroomest. 7.4 tok/s
- Mac Studio, M5 Max, 48GB$3,099Won't fit−6GB headroom
- Mac Studio, M5 Max, 64GB$3,499Fits8GB headroomest. 8.3 tok/s
- Fits70GB headroomest. 3.7 tok/s
- Fits38GB headroomest. 11.1 tok/s
- Fits69GB headroomest. 7.4 tok/s
- Fits70GB headroomest. 3.7 tok/s
- Fits69GB headroomest. 8.3 tok/s
- Fits38GB headroomest. 16.3 tok/s
- Fits197GB headroomest. 11.1 tok/s
- Fits197GB headroomest. 16.3 tok/s
- Fits453GB headroomest. 16.3 tok/s
| Config | Memory | Price | $/GB | Fit | Headroom | Est. speed |
|---|---|---|---|---|---|---|
| Mac mini, M6 | 16GB | $1,099 | $69 | Won't fit | −38GB | — |
| Mac mini, M6 | 24GB | $1,299 | $54 | Won't fit | −30GB | — |
| Mac mini, M6 | 32GB | $1,499 | $47 | Won't fit | −22GB | — |
| Mac mini, M5 Pro | 24GB | $1,699 | $71 | Won't fit | −30GB | — |
| Mac mini, M5 Pro | 48GB | $2,499 | $52 | Won't fit | −6GB | — |
| Mac Studio, M5 Max | 36GB | $2,499 | $69 | Won't fit | −18GB | — |
| Mac mini, M5 Pro | 64GB | $2,899 | $45 | Fits | 8GB | 4.2 tok/s |
| GeForce RTX 5090 | 32GB | $2,999.99+host | $94 | Won't fit | −20GB | — |
| Mac Studio, M4 Max, refurb | 64GB | $3,049 | $48 | Fits | 8GB | 7.4 tok/s |
| Mac Studio, M5 Max | 48GB | $3,099 | $65 | Won't fit | −6GB | — |
| Mac Studio, M5 Max | 64GB | $3,499 | $55 | Fits | 8GB | 8.3 tok/s |
| Ascent GX10, GB10 Grace Blackwell | 128GB | $3,999.99 | $31 | Fits | 70GB | 3.7 tok/s |
| Mac Studio, M3 Ultra, refurb | 96GB | $4,329 | $45 | Fits | 38GB | 11.1 tok/s |
| Mac Studio, M4 Max, refurb | 128GB | $4,409 | $34 | Fits | 69GB | 7.4 tok/s |
| DGX Spark, GB10 Grace Blackwell | 128GB | $4,699 | $37 | Fits | 70GB | 3.7 tok/s |
| Mac Studio, M5 Max | 128GB | $5,099 | $40 | Fits | 69GB | 8.3 tok/s |
| Mac Studio, M5 Ultra | 96GB | $5,499 | $57 | Fits | 38GB | 16.3 tok/s |
| Mac Studio, M3 Ultra, refurb | 256GB | $7,729 | $30 | Fits | 197GB | 11.1 tok/s |
| Mac Studio, M5 Ultra | 256GB | $9,499 | $37 | Fits | 197GB | 16.3 tok/s |
| Mac Studio, M5 Ultra | 512GB | — | — | Fits | 453GB | 16.3 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 = 24B × 2.000 B/param (16 bits) + 0.164 GB/1K × 8K K/V + 1.2 GB runtime
speed = bandwidth × 0.65 ÷ (24B active × 2.000 B/param) = 48.0 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 mini M5 Pro 64GB runs
- Llama-4-Scout-17B-16EQ3_K_M, est. 18.5 tok/s
- Llama-3.3-70BQ5_K_M, est. 4.0 tok/s
- Qwen3-32BQ8_0, est. 5.9 tok/s
- Qwen3-30B-A3BQ8_0, est. 43.8 tok/s
- Gemma-3-27BBF16, est. 3.7 tok/s
- Devstral-Small-2507BF16, est. 4.2 tok/s
Best quant that still fits at 8K context, largest model first. Full list on the Config page.
Bandwidth 307GB/s, usable memory 59GB of 64GB.