⌘K
← All AI Models

Can you run Step-5-Preview locally?

Step family · Sep 2026 · 604.3B parameters (27B activated per token) · Hugging Face model card

Step-5-Preview is a mixture-of-experts model released with open weights on Hugging Face: 604.34B total parameters (27.0B activated per token), a 1,048,576-token context window, other license. Running it locally needs roughly 384 GB of memory at Q4-class quantization (calculated) — multi-GPU server or large-unified-memory territory. Measured local speeds: not yet published here; we list unknowns as unknowns.

Minimum: 384 GB+ memory (Q4-class, calculated)  ·  Recommended: 512 GB+ for comfortable context headroom (Q4-class, calculated)

Parameters (total)
604.3B
Activated per token
27B (MoE)
Context window
1,048,576 tokens
Architecture source
HF config.json (verified)

How much VRAM does Step-5-Preview need at each quantization?

Step-5-Preview needs 398.9 GB of VRAM at Q4_K_M. The table below lists weights-only size and total VRAM including overhead for each common quantization level with 4k and 32k token contexts.

QuantizationBits / weightWeights onlyTotal + KV @4k ctxTotal + KV @32k ctx
Q4_K_M 4.8 362.6 GB 398.9 GB 398.9 GB
Q5_K_M 5.7 430.6 GB 473.7 GB 473.7 GB
Q6_K 6.6 498.6 GB 548.4 GB 548.4 GB
Q8_0 8.5 642.1 GB 706.3 GB 706.3 GB
FP16 16 1,208.7 GB 1,329.6 GB 1,329.6 GB

Method: weights = parameters × bits-per-weight (Q4_K_M ≈ 4.8, Q5_K_M ≈ 5.7, Q6_K ≈ 6.6, Q8_0 ≈ 8.5, FP16 = 16), plus 10% loading overhead, plus KV cache from the verified architecture config (not applicable to this architecture). Source: HF-verified 2026-10-13 via api (repo safetensors + config.json); active params from vendor README; VRAM figures are documented calculations, not measurements

Which GPUs can run Step-5-Preview locally?

At Q4_K_M with a 4k context, Step-5-Preview needs 398.9 GB of VRAM. The lists below are computed live from our GPU database and grouped by how much headroom the card has. Disclosure: CompareAIHardware.com participates in the Amazon Associates program and earns from qualifying purchases through links on this page. Affiliate relationships do not influence our recommendations.

Needs 2+ GPUs to run Step-5-Preview

One of these cards is too small on its own, but a pair (tensor or pipeline parallel, ~90% efficiency) covers the 398.9 GB requirement. See our multi-GPU guide for setup.

GPUVRAMBandwidthType
AMD Instinct MI325X 256 GB ×2 6000 GB/s Data Center GPU
AMD Instinct MI355X 288 GB ×2 8000 GB/s Data Center GPU
NVIDIA B300 (Blackwell Ultra) 288 GB ×2 8000 GB/s Data Center GPU

Frequently asked questions