Step 3.5 Flash NVFP4 — Hardware Requirements & GPU Compatibility
Chat
2.8K downloads
8 likesSpecifications
- Publisher
- tacos4me
- Parameters
- 111.3B
- Architecture
- Step3p5ForCausalLM
- Context Length
- 262,144 tokens
- Vocabulary Size
- 128,896
- Release Date
- 2026-02-21
- License
- Apache 2.0
Get Started
HuggingFace
How Much VRAM Does Step 3.5 Flash NVFP4 Need?
Select a quantization to see compatible GPUs below.
| Quantization | Bits | VRAM | + Context | File Size | Quality |
|---|---|---|---|---|---|
| IQ4_XS | 4.30 | 65.8 GB | — | 59.81 GB | Importance-weighted 4-bit, compact |
Which GPUs Can Run Step 3.5 Flash NVFP4?
IQ4_XS · 65.8 GBStep 3.5 Flash NVFP4 (IQ4_XS) requires 65.8 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 86+ GB is recommended. No single GPU has enough memory — multi-GPU or cluster setups are needed.
Which Devices Can Run Step 3.5 Flash NVFP4?
IQ4_XS · 65.8 GB5 devices with unified memory can run Step 3.5 Flash NVFP4, including NVIDIA DGX H100, NVIDIA DGX A100 640GB.
Related Models
Frequently Asked Questions
- How much VRAM does Step 3.5 Flash NVFP4 need?
Step 3.5 Flash NVFP4 requires 65.8 GB of VRAM at IQ4_XS.
VRAM = Weights + KV Cache + Overhead
Weights = 111.3B × 4.3 bits ÷ 8 = 59.8 GB
KV Cache + Overhead ≈ 6 GB (at 2K context + ~0.3 GB framework)
VRAM usage by quantization
IQ4_XS65.8 GB- Can NVIDIA GeForce RTX 5090 run Step 3.5 Flash NVFP4?
No — Step 3.5 Flash NVFP4 requires at least 65.8 GB at IQ4_XS, which exceeds the NVIDIA GeForce RTX 5090's 32 GB of VRAM.
- Can I run Step 3.5 Flash NVFP4 on a Mac?
Step 3.5 Flash NVFP4 requires at least 65.8 GB at IQ4_XS, which exceeds the unified memory of most consumer Macs. You would need a Mac Studio or Mac Pro with a high-memory configuration.
- Can I run Step 3.5 Flash NVFP4 locally?
Yes — Step 3.5 Flash NVFP4 can run locally on consumer hardware. At IQ4_XS quantization it needs 65.8 GB of VRAM. Popular tools include Ollama, LM Studio, and llama.cpp.
- How fast is Step 3.5 Flash NVFP4?
At IQ4_XS, Step 3.5 Flash NVFP4 can reach ~44 tok/s on AMD Instinct MI300X. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.
tok/s = (bandwidth GB/s ÷ model GB) × efficiency
Example: AMD Instinct MI300X → 5300 ÷ 65.8 × 0.55 = ~44 tok/s
Estimated speed at IQ4_XS (65.8 GB)
AMD Instinct MI300X~44 tok/sNVIDIA H100 SXM~33 tok/sAMD Instinct MI250X~27 tok/sReal-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
- What's the download size of Step 3.5 Flash NVFP4?
At IQ4_XS, the download is about 59.81 GB.
- Which GPUs can run Step 3.5 Flash NVFP4?
No single consumer GPU has enough VRAM to run Step 3.5 Flash NVFP4 at IQ4_XS (65.8 GB). Multi-GPU or professional hardware is required.
- Which devices can run Step 3.5 Flash NVFP4?
5 devices with unified memory can run Step 3.5 Flash NVFP4 at IQ4_XS (65.8 GB), including Mac Pro M2 Ultra (192 GB), Mac Studio M2 Ultra (192 GB), Mac Studio M4 Max (128 GB), NVIDIA DGX A100 640GB. Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.