Cheapest box to run Qwen3-VL-8B-Instruct at BF16
Alibaba’s Qwen3-VL-8B-Instruct is 8B 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.
Qwen3-VL-8B-Instruct at BF16 needs 18GB.
Cheapest Config that fits: Mac mini, M6, 32GB. Fits, 11GB headroom. Est. 6.9 tok/s.
Best offer today: $1,499 at Apple — in stock, checked 1 hour ago.
Collecting price history — 3 of 90 days. No timing verdict yet.
Offers — Mac mini, M6, 32GB, 512GB
- Mac mini, M6, 16GB$1,099Won't fit−5GB headroom
- Mac mini, M6, 24GB$1,299Tight3GB headroomest. 6.9 tok/s
- Mac mini, M6, 32GB$1,499Fits11GB headroomest. 6.9 tok/s
- Mac mini, M5 Pro, 24GB$1,699Tight3GB headroomest. 12.5 tok/s
- Mac mini, M5 Pro, 48GB$2,499Fits26GB headroomest. 12.5 tok/s
- Mac Studio, M5 Max, 36GB$2,499Fits15GB headroomest. 24.9 tok/s
- Mac mini, M5 Pro, 64GB$2,899Fits41GB headroomest. 12.5 tok/s
- GeForce RTX 5090, 32GB$2,999.99+hostFits12GB headroomest. 72.8 tok/s
- Fits41GB headroomest. 22.2 tok/s
- Mac Studio, M5 Max, 48GB$3,099Fits26GB headroomest. 24.9 tok/s
- Mac Studio, M5 Max, 64GB$3,499Fits41GB headroomest. 24.9 tok/s
- Fits102GB headroomest. 11.1 tok/s
- Fits70GB headroomest. 33.3 tok/s
- Fits102GB headroomest. 22.2 tok/s
- Fits102GB headroomest. 11.1 tok/s
- Fits102GB headroomest. 24.9 tok/s
- Fits70GB headroomest. 48.8 tok/s
- Fits230GB headroomest. 33.3 tok/s
- Fits230GB headroomest. 48.8 tok/s
- Fits486GB headroomest. 48.8 tok/s
| Config | Memory | Price | $/GB | Fit | Headroom | Est. speed |
|---|---|---|---|---|---|---|
| Mac mini, M6 | 16GB | $1,099 | $69 | Won't fit | −5GB | — |
| Mac mini, M6 | 24GB | $1,299 | $54 | Tight | 3GB | 6.9 tok/s |
| Mac mini, M6 | 32GB | $1,499 | $47 | Fits | 11GB | 6.9 tok/s |
| Mac mini, M5 Pro | 24GB | $1,699 | $71 | Tight | 3GB | 12.5 tok/s |
| Mac mini, M5 Pro | 48GB | $2,499 | $52 | Fits | 26GB | 12.5 tok/s |
| Mac Studio, M5 Max | 36GB | $2,499 | $69 | Fits | 15GB | 24.9 tok/s |
| Mac mini, M5 Pro | 64GB | $2,899 | $45 | Fits | 41GB | 12.5 tok/s |
| GeForce RTX 5090 | 32GB | $2,999.99+host | $94 | Fits | 12GB | 72.8 tok/s |
| Mac Studio, M4 Max, refurb | 64GB | $3,049 | $48 | Fits | 41GB | 22.2 tok/s |
| Mac Studio, M5 Max | 48GB | $3,099 | $65 | Fits | 26GB | 24.9 tok/s |
| Mac Studio, M5 Max | 64GB | $3,499 | $55 | Fits | 41GB | 24.9 tok/s |
| Ascent GX10, GB10 Grace Blackwell | 128GB | $3,999.99 | $31 | Fits | 102GB | 11.1 tok/s |
| Mac Studio, M3 Ultra, refurb | 96GB | $4,329 | $45 | Fits | 70GB | 33.3 tok/s |
| Mac Studio, M4 Max, refurb | 128GB | $4,409 | $34 | Fits | 102GB | 22.2 tok/s |
| DGX Spark, GB10 Grace Blackwell | 128GB | $4,699 | $37 | Fits | 102GB | 11.1 tok/s |
| Mac Studio, M5 Max | 128GB | $5,099 | $40 | Fits | 102GB | 24.9 tok/s |
| Mac Studio, M5 Ultra | 96GB | $5,499 | $57 | Fits | 70GB | 48.8 tok/s |
| Mac Studio, M3 Ultra, refurb | 256GB | $7,729 | $30 | Fits | 230GB | 33.3 tok/s |
| Mac Studio, M5 Ultra | 256GB | $9,499 | $37 | Fits | 230GB | 48.8 tok/s |
| Mac Studio, M5 Ultra | 512GB | — | — | Fits | 486GB | 48.8 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 = 8B × 2.000 B/param (16 bits) + 0.147 GB/1K × 8K K/V + 1.2 GB runtime
speed = bandwidth × 0.65 ÷ (8B active × 2.000 B/param) = 16.0 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 M6 32GB runs
- Qwen3-32BQ5_K_M, est. 4.9 tok/s
- Qwen3-30B-A3BQ6_K, est. 31.4 tok/s
- Gemma-3-27BQ6_K, est. 5.0 tok/s
- Devstral-Small-2507Q8_0, est. 4.3 tok/s
- gpt-oss-20bQ8_0, est. 22.2 tok/s
- Phi-4Q8_0, est. 7.4 tok/s
Best quant that still fits at 8K context, largest model first. Full list on the Config page.
Bandwidth 170GB/s, usable memory 29GB of 32GB.