What GPU do I need to run stepfun-ai/Step-3.5-Flash?
199.4B parameters, published in BF16. View on Hugging Face
Step-3.5-Flash is published by stepfun-ai on Hugging Face, with 155,341 downloads and 833 likes to date. It's a Step3p5ForCausalLM model built for text-generation, published natively in BF16.
VRAM required & cheapest live GPU fit
Required VRAM = weight size at each precision times a flat 1.2 overhead for activations and fragmentation. The KV-cache grows with context and is not in that factor; it is listed per model below where the architecture is published. Full formula and assumptions: methodology.
| Precision | Weight size | Required VRAM | Cheapest live fit | GPUs needed | Est. $/hr (full fit) |
|---|---|---|---|---|---|
| BF16 | 371.4 GB | 445.7 GB | RTX PRO 6000 | 5 | $5.98/hr |
| FP8 (quantized) | 185.7 GB | 222.8 GB | RTX PRO 6000 | 3 | $3.59/hr |
| INT4 (quantized) | 92.8 GB | 111.4 GB | AMD MI300X | 1 | $2.63/hr |
| cheaper alt. | RTX A5000 | 5 | $0.880/hr |
A GPU is only matched to a row if its hardware supports that precision, and the primary recommendation is always a single-GPU fit when one exists.
INT4 requires a quantized checkpoint actually published for this model, check its Hugging Face page before relying on this row.
Cheapest way to run Step-3.5-Flash at its published (BF16) precision: 5× RTX PRO 6000, at $1.20/hr per GPU ($5.98/hr total). Quantizing to FP8 or INT4 (rows above) can cost less, but requires a compatible quantized checkpoint to exist for this model.
Step-3.5-Flash: KV cache by context length
The KV cache is the memory the attention layers hold for every token of context, on top of the weights. It is not part of the flat 1.2x overhead in the table above, grows with context length and with every concurrent request, and is why a long-context deployment needs more VRAM than the table shows.
KV-cache size for Step-3.5-Flash is not published for this architecture, so no figure is shown.
Step-3.5-Flash: common questions
Can Step-3.5-Flash run on a single GPU?
No. At BF16 it needs 445.7 GB of VRAM, more than a single 32 GB desktop card holds. The cheapest capable card in the live feed is a 96.0 GB RTX PRO 6000, and it takes 5 of them.
How many GPUs do I need to run Step-3.5-Flash?
5 at BF16. It needs 445.7 GB of VRAM and the cheapest capable live offer is a 96.0 GB RTX PRO 6000, so 5 of them come to $5.98/hr in total.
Does quantizing Step-3.5-Flash lower the GPU bill?
Yes. At BF16 the cheapest live fit is 5 RTX PRO 6000 cards at $5.98/hr. At INT4 (quantized) it drops to one AMD MI300X at $2.63/hr, provided a quantized checkpoint exists for it.
Weight-to-VRAM math, the fit rules, and how live prices are normalized: full methodology.
More stepfun-ai models
- step3 (321.0B, BF16)
- GOT-OCR2_0 (716M, BF16)
- GOT-OCR-2.0-hf (561M, BF16)
Alternatives at this size
Other models for text-generation within about a third of Step-3.5-Flash's 199.4B parameters, from other model lines.
- MiniMax-M2.7 (228.7B, F8_E4M3)
- Qwen3-235B-A22B (235.1B, BF16)
- Hy3-preview (298.8B, BF16)
- DeepSeek-V2 (235.7B, BF16)
- Solar-Open2-250B (250.3B, BF16)
More on Step-3.5-Flash
Fits on a 288 GB GPU at FP8: every model that fits in 288 GB.
Fits on a 141 GB GPU at INT4: every model that fits in 141 GB.
Best chat and assistants models: how Step-3.5-Flash ranks against the rest.
Related reading: RTX PRO 6000 pricing and specs, How much VRAM you need for LLMs, Serving LLMs with vLLM, vLLM vs TensorRT-LLM vs SGLang, and Best GPU for LLM inference.
Your own GPU cloud accounts
One console for the GPU clouds you already rent from. Your keys stay on your machine.
$20 a month, flat