Recurrentgemma 9B — Hardware Requirements & GPU Compatibility
Chat
Recurrentgemma 9B is a 9.6B-parameter open language model from Google in the Gemma family. At BF16 it needs about 21.18 GB of VRAM — see which GPUs and Macs can run it below.
95 downloads
60 likesSpecifications
- Publisher
- Family
- Gemma
- Parameters
- 9.6B
- Release Date
- 2024-08-07
- License
- Gemma Terms
Get Started
HuggingFace
How Much VRAM Does Recurrentgemma 9B Need?
Select a quantization to see compatible GPUs below.
| Quantization | Bits | VRAM | + Context | File Size | Quality |
|---|---|---|---|---|---|
| BF16 | 16.00 | 21.2 GB | — | 19.26 GB | Brain floating point 16 — preferred for training |
Which GPUs Can Run Recurrentgemma 9B?
BF16 · 21.2 GBRecurrentgemma 9B (BF16) requires 21.2 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 28+ GB is recommended. 5 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti.
Runs great
— Plenty of headroom
Decent
— Enough VRAM, may be tightWhich Devices Can Run Recurrentgemma 9B?
BF16 · 21.2 GB21 devices with unified memory can run Recurrentgemma 9B, including NVIDIA DGX H100, NVIDIA DGX A100 640GB, Mac Mini M4 Pro (24 GB).
Runs great
— Plenty of headroomRelated Models
Frequently Asked Questions
- How much VRAM does Recurrentgemma 9B need?
Recurrentgemma 9B requires 21.2 GB of VRAM at BF16.
VRAM = Weights + KV Cache + Overhead
Weights = 9.6B × 16 bits ÷ 8 = 19.3 GB
KV Cache + Overhead ≈ 1.9 GB (at 2K context + ~0.3 GB framework)
VRAM usage by quantization
BF1621.2 GB- Can I run Recurrentgemma 9B on a Mac?
Recurrentgemma 9B requires at least 21.2 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 Recurrentgemma 9B locally?
Yes — Recurrentgemma 9B can run locally on consumer hardware. At BF16 quantization it needs 21.2 GB of VRAM. Popular tools include Ollama, LM Studio, and llama.cpp.
- How fast is Recurrentgemma 9B?
At BF16, Recurrentgemma 9B can reach ~138 tok/s on AMD Instinct MI300X. On NVIDIA GeForce RTX 4090: ~31 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 ÷ 21.2 × 0.55 = ~138 tok/s
Estimated speed at BF16 (21.2 GB)
~138 tok/s~31 tok/s~103 tok/s~85 tok/sReal-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
- What's the download size of Recurrentgemma 9B?
At BF16, the download is about 19.26 GB.
- Which GPUs can run Recurrentgemma 9B?
5 consumer GPUs can run Recurrentgemma 9B at BF16 (21.2 GB). Top options include NVIDIA GeForce RTX 5090, AMD Radeon RX 7900 XTX, NVIDIA GeForce RTX 3090. 1 GPU have plenty of headroom for comfortable inference.
- Which devices can run Recurrentgemma 9B?
21 devices with unified memory can run Recurrentgemma 9B at BF16 (21.2 GB), including Mac Mini M4 (32 GB), Mac Mini M4 Pro (24 GB), Mac Mini M4 Pro (48 GB), Mac Pro M2 Ultra (192 GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.