Skip to content
can-i-run-this-llm

Can I run this LLM?

Can a NVIDIA RTX Pro 6000 Blackwell (96GB) run Step 3.7 Flash (MoE)?

Step · Apache 2.0 · estimated for a 96.0 GB machine at 4K context.

Yes — it runs great.

Runs great

Best fit: Step 3.7 Flash (MoE) at Q2_K · 71.2 GB weights · 169–254 tok/s on a NVIDIA RTX Pro 6000 Blackwell (96GB).

Arena Text Overall · style-controlled

UnratedNo exact match

No reviewed exact Arena match

No reviewed exact model/configuration match in the Arena Text Overall style-controlled snapshot. No family or editorial score is substituted. Local quantization quality is unmeasured.

Snapshot published 2026-10-08 · snapshot imported 2026-10-11 UTC · validated last-good data · manual refresh

Arena leaderboard dataset · Arena / lmarena-ai, Arena Leaderboard Dataset (CC BY 4.0). Selected Text Overall style-controlled rows; field names adapted, ratings unchanged. · dataset CC BY 4.0 · revision e16f90232986

Total memory at 4K context
72.3 GB
Weights + KV cache + runtime overhead
Estimated generation
169–254 tok/s
Estimated first token
~1.1s

Every quantization on this machine

On small screens, each quant lists every metric below its name. On wider screens, scroll sideways for total memory, speed and fit. Keyboard: focus the table, then use the arrow keys. Quant names stay visible.

Step 3.7 Flash (MoE) quantizations on NVIDIA RTX Pro 6000 Blackwell (96GB), estimated at 4K context. Weights and total memory are shown separately.
QuantWeightsTotal memorySpeedFit
Q2_K71.2 GB72.3 GB169–254 tok/sRuns great
Q3_K_M91.8 GB92.9 GB— not residentWon't fit
Q4_K_M121.6 GB122.7 GB— not residentWon't fit
Q5_K_M142.0 GB143.1 GB— not residentWon't fit
Q6_K171.8 GB172.8 GB— not residentWon't fit
Q8_0212.0 GB213.1 GB— not residentWon't fit
Open the full calculator →Everything a NVIDIA RTX Pro 6000 Blackwell (96GB) runsStep 3.7 Flash (MoE) on other machinesWhat to buy to run it →

Ran Step 3.7 Flash (MoE) on a NVIDIA RTX Pro 6000 Blackwell (96GB)? Add your real numbers →