Zira Researcher — Hardware Requirements & GPU Compatibility
ChatReasoning
Zira Researcher is a 4.2B-parameter open language model from 0xvoid0000. It supports a context window of up to 262,144 tokens. At BF16 it needs about 8.88 GB of VRAM — see which GPUs and Macs can run it below.
57 downloads
3 likesSpecifications
- Publisher
- 0xvoid0000
- Parameters
- 4.2B
- Architecture
- Qwen3_5ForCausalLM
- Context Length
- 262,144 tokens
- Vocabulary Size
- 248,320
- Release Date
- 2026-03-14
- License
- Apache 2.0
Get Started
HuggingFace
How Much VRAM Does Zira Researcher Need?
Select a quantization to see compatible GPUs below.
| Quantization | Bits | VRAM | + Context | File Size | Quality |
|---|---|---|---|---|---|
| BF16est. | 16.00 | 8.9 GB | 30.2 GB | 8.41 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 Zira Researcher?
BF16 · 8.9 GBZira Researcher (BF16) requires 8.9 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 12+ GB is recommended. Using the full 262K context window can add up to 21.3 GB, bringing total usage to 30.2 GB. 28 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti, NVIDIA GeForce RTX 3080 Ti.
Runs great
— Plenty of headroom
Decent
— Enough VRAM, may be tightWhich Devices Can Run Zira Researcher?
BF16 · 8.9 GB27 devices with unified memory can run Zira Researcher, including NVIDIA DGX H100, NVIDIA DGX A100 640GB.
Runs great
— Plenty of headroomRelated Models
Frequently Asked Questions
- How much VRAM does Zira Researcher need?
Zira Researcher requires 8.9 GB of VRAM at BF16. Full 262K context adds up to 21.3 GB (30.2 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 4.2B × 16 bits ÷ 8 = 8.4 GB
KV Cache + Overhead ≈ 0.5 GB (at 2K context + ~0.3 GB framework)
KV Cache + Overhead ≈ 21.8 GB (at full 262K context)
VRAM usage by quantization
BF168.9 GBBF16 + full context30.2 GB- Can I run Zira Researcher on a Mac?
Zira Researcher requires at least 8.9 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 Zira Researcher locally?
Yes — Zira Researcher can run locally on consumer hardware. At BF16 quantization it needs 8.9 GB of VRAM. Popular tools include Ollama, LM Studio, and llama.cpp.
- How fast is Zira Researcher?
At BF16, Zira Researcher can reach ~328 tok/s on AMD Instinct MI300X. On NVIDIA GeForce RTX 4090: ~74 tok/s. 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 ÷ 8.9 × 0.55 = ~328 tok/s
Estimated speed at BF16 (8.9 GB)
~328 tok/s~74 tok/s~245 tok/s~203 tok/sReal-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
- What's the download size of Zira Researcher?
At BF16, the download is about 8.41 GB.
- Which GPUs can run Zira Researcher?
28 consumer GPUs can run Zira Researcher at BF16 (8.9 GB). Top options include AMD Radeon RX 6800, AMD Radeon RX 6800 XT, AMD Radeon RX 6900 XT, AMD Radeon RX 6700 XT. 17 GPUs have plenty of headroom for comfortable inference.
- Which devices can run Zira Researcher?
27 devices with unified memory can run Zira Researcher at BF16 (8.9 GB), including Mac Mini M4 (16 GB), Mac Mini M4 (32 GB), Mac Mini M4 Pro (24 GB), Mac Mini M4 Pro (48 GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.