CYBER FROST 3.8 EXL3 SAGE 3.87bpw — Hardware Requirements & GPU Compatibility
ChatFunctions
CYBER FROST 3.8 EXL3 SAGE 3.87bpw is a 32.8B-parameter open language model from vcruz305. It supports a context window of up to 262,144 tokens. At BF16 it needs about 66.07 GB of VRAM — see which GPUs and Macs can run it below.
1.3K downloads
45 likesSpecifications
- Publisher
- vcruz305
- Parameters
- 32.8B
- Architecture
- Qwen4ExpForConditionalGeneration
- Context Length
- 262,144 tokens
- Vocabulary Size
- 248,320
- Release Date
- 2026-09-28
- License
- Other
Get Started
HuggingFace
Run in cloud
Fits on A100 80GB (13 GB headroom) · BF16
- Generation speed
- ~20 tok/s
- generation speed
- Cost per 1M output tokens
- $14.96
- per 1M output tokens
How Much VRAM Does CYBER FROST 3.8 EXL3 SAGE 3.87bpw Need?
Select a quantization to see compatible GPUs below.
| Quantization | Bits | VRAM | + Context | File Size | Quality |
|---|---|---|---|---|---|
| BF16est. | 16.00 | 66.1 GB | 76.7 GB | 65.69 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 CYBER FROST 3.8 EXL3 SAGE 3.87bpw?
BF16 · 66.1 GBCYBER FROST 3.8 EXL3 SAGE 3.87bpw (BF16) requires 66.1 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 86+ GB is recommended. Using the full 262K context window can add up to 10.7 GB, bringing total usage to 76.7 GB. No consumer GPU has enough memory.
Which Devices Can Run CYBER FROST 3.8 EXL3 SAGE 3.87bpw?
BF16 · 66.1 GB19 devices with unified memory can run CYBER FROST 3.8 EXL3 SAGE 3.87bpw, including NVIDIA DGX H100, NVIDIA DGX A100 640GB, Mac Studio (M3 Ultra, 96GB).
Runs great
— Plenty of headroom
Decent
— Enough memory, may be tightRelated Models
Frequently Asked Questions
- How much VRAM does CYBER FROST 3.8 EXL3 SAGE 3.87bpw need?
CYBER FROST 3.8 EXL3 SAGE 3.87bpw requires 66.1 GB of VRAM at BF16. Full 262K context adds up to 10.7 GB (76.7 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 32.8B × 16 bits ÷ 8 = 65.7 GB
KV Cache + Overhead ≈ 0.4 GB (at 2K context + ~0.3 GB framework)
Fit ratings and hardware model lists check this model with room for a 16K-token context, which needs a little more memory.
KV Cache + Overhead ≈ 11 GB (at full 262K context)
VRAM usage by quantization
BF1666.1 GBBF16 + full context76.7 GB- Can NVIDIA GeForce RTX 5090 run CYBER FROST 3.8 EXL3 SAGE 3.87bpw?
No — CYBER FROST 3.8 EXL3 SAGE 3.87bpw requires at least 66.1 GB at BF16, which exceeds the NVIDIA GeForce RTX 5090's 32 GB of VRAM.
- Can I run CYBER FROST 3.8 EXL3 SAGE 3.87bpw on a Mac?
Yes — Mac Studio (M3 Ultra, 96GB) and 6 other Macs can run CYBER FROST 3.8 EXL3 SAGE 3.87bpw. Apple Silicon uses unified memory, so the model shares RAM with the system. At BF16 you need at least 66.1 GB of usable unified memory (RAM minus macOS overhead).
- Can I run CYBER FROST 3.8 EXL3 SAGE 3.87bpw locally?
Yes — CYBER FROST 3.8 EXL3 SAGE 3.87bpw can run locally on consumer hardware. At BF16 quantization it needs 66.1 GB of VRAM. Popular tools include Ollama, LM Studio, and llama.cpp.
- How fast is CYBER FROST 3.8 EXL3 SAGE 3.87bpw?
At BF16, CYBER FROST 3.8 EXL3 SAGE 3.87bpw can reach ~73 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 B200 → 8000 ÷ 66.1 × 0.65 = ~79 tok/s
Estimated speed at BF16 (66.1 GB)
~79 tok/s~79 tok/s~73 tok/sReal-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
- What's the download size of CYBER FROST 3.8 EXL3 SAGE 3.87bpw?
At BF16, the download is about 65.69 GB.
- Which GPUs can run CYBER FROST 3.8 EXL3 SAGE 3.87bpw?
No single consumer GPU has enough VRAM to run CYBER FROST 3.8 EXL3 SAGE 3.87bpw at BF16 (66.1 GB). Multi-GPU or professional hardware is required.
- Which devices can run CYBER FROST 3.8 EXL3 SAGE 3.87bpw?
19 devices with unified memory can run CYBER FROST 3.8 EXL3 SAGE 3.87bpw at BF16 (66.1 GB), including ASUS Ascent GX10, Asus ROG Flow Z13 (2025, Ryzen AI Max+ 395, 128 GB), Beelink GTR9 Pro (Ryzen AI Max+ 395, 128 GB), Framework Desktop (Ryzen AI Max+ 395, 128 GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.