Cheapest box to run DeepSeek-V4-Flash at Q8_0
DeepSeek’s DeepSeek-V4-Flash is 284B parameters, 13B 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.
DeepSeek-V4-Flash at Q8_0 needs 303GB.
Cheapest Config that fits: Mac Studio, M5 Ultra, 512GB. Fits, 201GB headroom. Est. 43.4 tok/s.
No price yet: Ships late October 2026. Apple has not published a price.
Price history: 34 days recorded.
Compressed sparse attention. Each layer stores one latent, divided by the compression ratio in its config (mostly 4x or 128x); the figure matches DeepSeek's own claim of about a tenth of the V3 cache.
Offers — Mac Studio, M5 Ultra, 512GB, 1TB
Ships late October 2026. Apple has not published a price.
- Mac mini, M6, 16GB$1,099Won't fit−290GB headroom
- Mac mini, M6, 24GB$1,299Won't fit−282GB headroom
- Mac mini, M6, 32GB$1,499Won't fit−274GB headroom
- Mac mini, M5 Pro, 24GB$1,699Won't fit−282GB headroom
- Mac mini, M5 Pro, 48GB$2,499Won't fit−259GB headroom
- Mac Studio, M5 Max, 36GB$2,499Won't fit−270GB headroom
- Mac mini, M5 Pro, 64GB$2,899Won't fit−244GB headroom
- GeForce RTX 5090, 32GB$2,999.99+hostWon't fit−272GB headroom
- Won't fit−244GB headroom
- Mac Studio, M5 Max, 48GB$3,099Won't fit−259GB headroom
- Mac Studio, M5 Max, 64GB$3,499Won't fit−244GB headroom
- Won't fit−183GB headroom
- Won't fit−215GB headroom
- Won't fit−183GB headroom
- Won't fit−183GB headroom
- Won't fit−183GB headroom
- Won't fit−215GB headroom
- Won't fit−55GB headroom
- Won't fit−55GB headroom
- Fits201GB headroomest. 43.4 tok/s
| Config | Memory | Price | $/GB | Fit | Headroom | Est. speed |
|---|---|---|---|---|---|---|
| Mac mini, M6 | 16GB | $1,099 | $69 | Won't fit | −290GB | — |
| Mac mini, M6 | 24GB | $1,299 | $54 | Won't fit | −282GB | — |
| Mac mini, M6 | 32GB | $1,499 | $47 | Won't fit | −274GB | — |
| Mac mini, M5 Pro | 24GB | $1,699 | $71 | Won't fit | −282GB | — |
| Mac mini, M5 Pro | 48GB | $2,499 | $52 | Won't fit | −259GB | — |
| Mac Studio, M5 Max | 36GB | $2,499 | $69 | Won't fit | −270GB | — |
| Mac mini, M5 Pro | 64GB | $2,899 | $45 | Won't fit | −244GB | — |
| GeForce RTX 5090 | 32GB | $2,999.99+host | $94 | Won't fit | −272GB | — |
| Mac Studio, M4 Max, refurb | 64GB | $3,049 | $48 | Won't fit | −244GB | — |
| Mac Studio, M5 Max | 48GB | $3,099 | $65 | Won't fit | −259GB | — |
| Mac Studio, M5 Max | 64GB | $3,499 | $55 | Won't fit | −244GB | — |
| Ascent GX10, GB10 Grace Blackwell | 128GB | $3,999.99 | $31 | Won't fit | −183GB | — |
| Mac Studio, M3 Ultra, refurb | 96GB | $4,329 | $45 | Won't fit | −215GB | — |
| Mac Studio, M4 Max, refurb | 128GB | $4,409 | $34 | Won't fit | −183GB | — |
| DGX Spark, GB10 Grace Blackwell | 128GB | $4,699 | $37 | Won't fit | −183GB | — |
| Mac Studio, M5 Max | 128GB | $5,099 | $40 | Won't fit | −183GB | — |
| Mac Studio, M5 Ultra | 96GB | $5,499 | $57 | Won't fit | −215GB | — |
| Mac Studio, M3 Ultra, refurb | 256GB | $7,729 | $30 | Won't fit | −55GB | — |
| Mac Studio, M5 Ultra | 256GB | $9,499 | $37 | Won't fit | −55GB | — |
| Mac Studio, M5 Ultra | 512GB | — | — | Fits | 201GB | 43.4 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 = 284B × 1.063 B/param (8.5 bits) + 0.006 GB/1K × 8K K/V + 1.2 GB runtime
speed = bandwidth × 0.5 ÷ (13B active × 1.063 B/param) = 13.8 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 512GB runs
- Kimi-K2-InstructQ3_K_M, est. 38.4 tok/s
- DeepSeek-V3Q5_K_M, est. 22.9 tok/s
- DeepSeek-R1Q5_K_M, est. 22.9 tok/s
- Qwen3-Coder-480B-A35B-InstructQ6_K, est. 20.9 tok/s
- Llama-4-Maverick-17B-128EQ8_0, est. 33.2 tok/s
- GLM-4.6Q8_0, est. 17.6 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 504GB of 512GB.