Tenebra 30B Alpha01 — Hardware Requirements & GPU Compatibility
Chat
Tenebra 30B Alpha01 is a 32.5B-parameter open language model from SicariusSicariiStuff. It supports a context window of up to 16,384 tokens. At BF16 it needs about 68.63 GB of VRAM — see which GPUs and Macs can run it below.
19 downloads
16 likesSpecifications
- Publisher
- SicariusSicariiStuff
- Parameters
- 32.5B
- Architecture
- LlamaForCausalLM
- Context Length
- 16,384 tokens
- Vocabulary Size
- 32,000
- Release Date
- 2025-11-05
Get Started
HuggingFace
How Much VRAM Does Tenebra 30B Alpha01 Need?
Select a quantization to see compatible GPUs below.
| Quantization | Bits | VRAM | + Context | File Size | Quality |
|---|---|---|---|---|---|
| BF16 | 16.00 | 68.6 GB | 91.5 GB | 65.06 GB | Brain floating point 16 — preferred for training |
Which GPUs Can Run Tenebra 30B Alpha01?
BF16 · 68.6 GBTenebra 30B Alpha01 (BF16) requires 68.6 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 90+ GB is recommended. Using the full 16K context window can add up to 22.9 GB, bringing total usage to 91.5 GB. No single GPU has enough memory — multi-GPU or cluster setups are needed.
Which Devices Can Run Tenebra 30B Alpha01?
BF16 · 68.6 GB5 devices with unified memory can run Tenebra 30B Alpha01, including NVIDIA DGX H100, NVIDIA DGX A100 640GB.
Related Models
Frequently Asked Questions
- How much VRAM does Tenebra 30B Alpha01 need?
Tenebra 30B Alpha01 requires 68.6 GB of VRAM at BF16. Full 16K context adds up to 22.9 GB (91.5 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 32.5B × 16 bits ÷ 8 = 65.1 GB
KV Cache + Overhead ≈ 3.5 GB (at 2K context + ~0.3 GB framework)
KV Cache + Overhead ≈ 26.4 GB (at full 16K context)
VRAM usage by quantization
BF1668.6 GBBF16 + full context91.5 GB- Can NVIDIA GeForce RTX 5090 run Tenebra 30B Alpha01?
No — Tenebra 30B Alpha01 requires at least 68.6 GB at BF16, which exceeds the NVIDIA GeForce RTX 5090's 32 GB of VRAM.
- Can I run Tenebra 30B Alpha01 on a Mac?
Tenebra 30B Alpha01 requires at least 68.6 GB at BF16, 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 Tenebra 30B Alpha01 locally?
Yes — Tenebra 30B Alpha01 can run locally on consumer hardware. At BF16 quantization it needs 68.6 GB of VRAM. Popular tools include Ollama, LM Studio, and llama.cpp.
- How fast is Tenebra 30B Alpha01?
At BF16, Tenebra 30B Alpha01 can reach ~43 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 ÷ 68.6 × 0.55 = ~43 tok/s
Estimated speed at BF16 (68.6 GB)
~43 tok/s~32 tok/s~26 tok/sReal-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
- What's the download size of Tenebra 30B Alpha01?
At BF16, the download is about 65.06 GB.
- Which GPUs can run Tenebra 30B Alpha01?
No single consumer GPU has enough VRAM to run Tenebra 30B Alpha01 at BF16 (68.6 GB). Multi-GPU or professional hardware is required.
- Which devices can run Tenebra 30B Alpha01?
5 devices with unified memory can run Tenebra 30B Alpha01 at BF16 (68.6 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.