Step3 — Hardware Requirements & GPU Compatibility
Vision
Step3 is StepFun's flagship multimodal reasoning model, a mixture-of-experts vision-language system with 321 billion total parameters (316 billion in the language backbone) and about 38 billion active per token, routed across 48 experts with 3 selected plus 1 shared expert per token. It combines two co-designed techniques, Multi-Matrix Factorization Attention and Attention-FFN Disaggregation, specifically to cut decoding cost so the model stays efficient on both flagship and lower-end accelerators despite its size. Even with that efficiency focus, its total footprint still calls for a multi-GPU server rather than a single workstation card. Context length is 65,536 tokens. It is released under the Apache 2.0 license, permitting unrestricted commercial and research use, and was published in July 2025.
39.7K downloads
166 likesSpecifications
- Publisher
- StepFun
- Family
- Step 3
- Parameters
- 321.0B
- Architecture
- Step3VLForConditionalGeneration
- Context Length
- 65,536 tokens
- Vocabulary Size
- 128,815
- Release Date
- 2025-07-28
- License
- Apache 2.0
Get Started
HuggingFace
How Much VRAM Does Step3 Need?
Select a quantization to see compatible GPUs below.
| Quantization | Bits | VRAM | + Context | File Size | Quality |
|---|---|---|---|---|---|
| Q2_Kest. | 3.40 | 150.1 GB | — | 136.41 GB | 2-bit quantization with K-quant improvements |
| Q3_K_Mest. | 3.90 | 172.1 GB | — | 156.47 GB | 3-bit medium quantization |
| Q4_K_Mest. | 4.80 | 211.8 GB | — | 192.58 GB | 4-bit medium quantization — most popular sweet spot |
| Q5_K_Mest. | 5.70 | 251.6 GB | — | 228.69 GB | 5-bit medium quantization — good quality/size tradeoff |
| Q6_Kest. | 6.60 | 291.3 GB | — | 264.80 GB | 6-bit quantization, very good quality |
| Q8_0est. | 8.00 | 353.1 GB | — | 320.97 GB | 8-bit quantization, near-lossless |
| BF16est. | 16.00 | 706.1 GB | — | 641.94 GB | Brain floating point 16 — preferred for training |
est.= calculated VRAM estimate; no published GGUF file found for that quantization yet. Other rows are verified against real community uploads.
Which GPUs Can Run Step3?
Q4_K_M · 211.8 GBStep3 (Q4_K_M) requires 211.8 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 276+ GB is recommended. No single GPU has enough memory — multi-GPU or cluster setups are needed.
Which Devices Can Run Step3?
Q4_K_M · 211.8 GB3 devices with unified memory can run Step3, including NVIDIA DGX H100, NVIDIA DGX A100 640GB.
Runs great
— Plenty of headroomBenchmarks
Benchmark details →Related Models
Frequently Asked Questions
- How much VRAM does Step3 need?
Step3 requires 211.8 GB of VRAM at Q4_K_M, or 706.1 GB at BF16.
VRAM = Weights + KV Cache + Overhead
Weights = 321.0B × 4.8 bits ÷ 8 = 192.6 GB
KV Cache + Overhead ≈ 19.2 GB (at 2K context + ~0.3 GB framework)
VRAM usage by quantization
Q4_K_M211.8 GB- Can NVIDIA GeForce RTX 5090 run Step3?
No — Step3 requires at least 150.1 GB at Q2_K, which exceeds the NVIDIA GeForce RTX 5090's 32 GB of VRAM.
- What's the best quantization for Step3?
For Step3, Q4_K_M (211.8 GB) offers the best balance of quality and VRAM usage. Q5_K_M (251.6 GB) provides better quality if you have the VRAM. The smallest option is Q2_K at 150.1 GB.
VRAM requirement by quantization
Q2_K150.1 GBQ4_K_M ★211.8 GBQ5_K_M251.6 GBQ6_K291.3 GBQ8_0353.1 GBBF16706.1 GB★ Recommended — best balance of quality and VRAM usage.
- Can I run Step3 on a Mac?
Step3 requires at least 150.1 GB at Q2_K, 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 Step3 locally?
Yes — Step3 can run locally on consumer hardware. At Q4_K_M quantization it needs 211.8 GB of VRAM. Popular tools include Ollama, LM Studio, and llama.cpp.
- How fast is Step3?
At Q4_K_M, Step3 can reach ~23 tok/s on AMD Instinct MI350X. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.
tok/s = (bandwidth GB/s ÷ model GB) × efficiency
Example: NVIDIA B300 → 8000 ÷ 211.8 × 0.65 = ~25 tok/s
Estimated speed at Q4_K_M (211.8 GB)
~25 tok/s~23 tok/s~23 tok/sReal-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
- What's the download size of Step3?
At Q4_K_M, the download is about 192.58 GB. The full-precision BF16 version is 641.94 GB. The smallest option (Q2_K) is 136.41 GB.
- Which GPUs can run Step3?
No single consumer GPU has enough VRAM to run Step3 at Q4_K_M (211.8 GB). Multi-GPU or professional hardware is required.
- Which devices can run Step3?
4 devices with unified memory can run Step3 at Q4_K_M (211.8 GB), including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB), NVIDIA DGX A100 640GB, NVIDIA DGX H100. Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.