Cheapest box to run gpt-oss-120b at Q4_K_M
OpenAI’s gpt-oss-120b is 117B parameters, 5.1B 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.
gpt-oss-120b at Q4_K_M needs 72GB.
Cheapest Config that fits: DGX Spark, GB10 Grace Blackwell, 128GB. Fits, 48GB headroom. Est. 44.3 tok/s.
Best offer today: $4,699 at NVIDIA — in stock, checked 9 hours ago.
Collecting price history — 1 of 90 days. No timing verdict yet.
Offers — DGX Spark, GB10 Grace Blackwell, 128GB, 4TB
- NVIDIAlist pricein stockchecked 9 hours ago$4,699
- Mac mini, M6, 16GB$1,099Won't fit−59GB headroom
- Mac mini, M6, 24GB$1,299Won't fit−51GB headroom
- Mac mini, M6, 32GB$1,499Won't fit−43GB headroom
- Mac mini, M5 Pro, 24GB$1,699Won't fit−51GB headroom
- Mac mini, M5 Pro, 48GB$2,499Won't fit−28GB headroom
- Mac Studio, M5 Max, 36GB$2,499Won't fit−39GB headroom
- Mac mini, M5 Pro, 64GB$2,899Won't fit−14GB headroom
- Mac Studio, M5 Max, 48GB$3,099Won't fit−28GB headroom
- Mac Studio, M5 Max, 64GB$3,499Won't fit−14GB headroom
- Fits48GB headroomest. 44.3 tok/s
- Fits48GB headroomest. 99.7 tok/s
- Fits16GB headroomest. 195 tok/s
- GeForce RTX 5090, 32GB$5,699.99+hostWon't fit−42GB headroom
- Fits176GB headroomest. 195 tok/s
- Fits432GB headroomest. 195 tok/s
| Config | Memory | Price | $/GB | Fit | Headroom | Est. speed |
|---|---|---|---|---|---|---|
| Mac mini, M6 | 16GB | $1,099 | $69 | Won't fit | −59GB | — |
| Mac mini, M6 | 24GB | $1,299 | $54 | Won't fit | −51GB | — |
| Mac mini, M6 | 32GB | $1,499 | $47 | Won't fit | −43GB | — |
| Mac mini, M5 Pro | 24GB | $1,699 | $71 | Won't fit | −51GB | — |
| Mac mini, M5 Pro | 48GB | $2,499 | $52 | Won't fit | −28GB | — |
| Mac Studio, M5 Max | 36GB | $2,499 | $69 | Won't fit | −39GB | — |
| Mac mini, M5 Pro | 64GB | $2,899 | $45 | Won't fit | −14GB | — |
| Mac Studio, M5 Max | 48GB | $3,099 | $65 | Won't fit | −28GB | — |
| Mac Studio, M5 Max | 64GB | $3,499 | $55 | Won't fit | −14GB | — |
| DGX Spark, GB10 Grace Blackwell | 128GB | $4,699 | $37 | Fits | 48GB | 44.3 tok/s |
| Mac Studio, M5 Max | 128GB | $5,099 | $40 | Fits | 48GB | 99.7 tok/s |
| Mac Studio, M5 Ultra | 96GB | $5,499 | $57 | Fits | 16GB | 195 tok/s |
| GeForce RTX 5090 | 32GB | $5,699.99+host | $178 | Won't fit | −42GB | — |
| Mac Studio, M5 Ultra | 256GB | $9,499 | $37 | Fits | 176GB | 195 tok/s |
| Mac Studio, M5 Ultra | 512GB | — | — | Fits | 432GB | 195 tok/s |
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 = 117B × 0.604 B/param (4.83 bits) + 0.074 GB/1K × 8K K/V + 1.2 GB runtime
speed = bandwidth × 0.5 ÷ (5.1B active × 0.604 B/param) = 3.1 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 DGX Spark GB10 Grace Blackwell 128GB runs
- Qwen3-235B-A22BQ3_K_M, est. 12.7 tok/s
- Mistral-Large-2411Q6_K, est. 1.8 tok/s
- gpt-oss-120bQ6_K, est. 32.6 tok/s
- Llama-4-Scout-17B-16EQ8_0, est. 7.6 tok/s
- Llama-3.3-70BQ8_0, est. 2.4 tok/s
- Qwen3-32BBF16, est. 2.8 tok/s
Best quant that still fits at 8K context, largest model first. Full list on the Config page.
Bandwidth 273GB/s, usable memory 120GB of 128GB.