Japanese GPT Neox Small — Hardware Requirements & GPU Compatibility
Chat
Japanese GPT Neox Small is a 204M-parameter open language model from rinna. It supports a context window of up to 2,048 tokens. At Q4_K_M it needs about 0.13 GB of VRAM — see which GPUs and Macs can run it below.
Runs great on a 24 GB GPU
Runs from 8 GB (Q4_K_M · 2K)Best on 24 GB (Q4_K_M · ~5040 tok/s · 2K)Macs from 8 GB (Q4_K_M · ~551 tok/s)
MIT
Reference GPU · RTX 4090 24 GB
Estimates, ±20% · how we estimate
Run Japanese GPT Neox Small
One recommended way. Every other method, OS and quant is in the step-by-step guide.Each way opens a step-by-step guide for your OS and quant.
Recommended
vllm serve rinna/japanese-gpt-neox-small --max-model-len 2048vLLM repo with its own quantisation
Download: Weights on Hugging Face
Ways to runOther ways
each opens step-by-stepRun in cloud
Runs on RTX 3080 Ti (12 GB) (11 GB headroom) · Q4_K_M
- Generation speed
- ~4562 tok/s
- generation speed
- Cost per 1M output tokens
- $0.0074
- per 1M output tokens
How Much VRAM Does Japanese GPT Neox Small Need?
Select a quantization level; the listed hardware adheres to it.
| Quantization | VRAM | +32K | On RTX 4090 (reference) | |
|---|---|---|---|---|
| 0.1 GB | — | 2-bit · K-quant — 2-bit quantization with K-quant improvements | Fits: 2K context · ~6552 tok/s | |
| 0.1 GB | — | 3-bit medium — 3-bit medium quantization | Fits: 2K context · ~5956 tok/s | |
| 0.1 GB | — | 4-bit medium — 4-bit medium quantization — most popular sweet spot | Fits: 2K context · ~5040 tok/s | |
| 0.2 GB | — | 5-bit medium — 5-bit medium quantization — good quality/size tradeoff | Fits: 2K context · ~4095 tok/s | |
| 0.2 GB | — | 6-bit, very good — 6-bit quantization, very good quality | Fits: 2K context · ~3640 tok/s | |
| 0.2 GB | — | 8-bit, near-lossless — 8-bit quantization, near-lossless | Fits: 2K context · ~2978 tok/s | |
| 0.5 GB | — | 16-bit · bfloat — Brain floating point 16 — preferred for training | Fits: 2K context · ~1456 tok/s |
Selected Q4_K_M
4-bit medium quantization0.1GBat 2K context
- Weights
- 0.1 GB
- KV cache
- 0.0 GB
- Overhead
- 0.3 GB
Fits RTX 4090 (reference) · 23.9 GB free · ~5040 tok/s
≥50 tok/s needs ≥10 GB/s · RTX 5090 ~8960, RTX 3090 Ti ~5040
Which Hardware Can Run Japanese GPT Neox Small?
Q4_K_M · 0.1 GBJapanese GPT Neox Small (Q4_K_M) requires 0.1 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 1+ GB is recommended. 52 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti. 163 devices with unified memory can run Japanese GPT Neox Small, including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB).
≤ 11 GB
≤ 11 GB tier →fits · 7.9 GB free · up to 2K ctx
16 GPUs · 1 Mac class · 2 device families
RTX 3080~3802Simulate output on RTX 3080RTX 2080 Ti~3080Simulate output on RTX 2080 TiRTX 3070 Ti~3042Simulate output on RTX 3070 TiM3*8 GB~551Simulate output on iMac 24" M3 (8 GB)Orin Nano~510Simulate output on NVIDIA Jetson Orin Nano 8GB (Super)iPhone~367Simulate output on iPhone 1720 GB
fits · 19.9 GB free · up to 2K ctx
1 GPU
RX 7900 XT~3692Simulate output on RX 7900 XT32–40 GB
fits · 31.9 GB free · up to 2K ctx
1 GPU · 3 Mac classes · 4 device families
RTX 5090~8960Simulate output on RTX 5090M5 Max36 GB~2477Simulate output on Mac Studio M5 Max (36 GB)M532 GB~827Simulate output on MacBook Air 13" M5 (32 GB)Orin~1024Simulate output on NVIDIA Jetson AGX Orin 32GBSnapdragon X Elite (X1E-80-100)~675Simulate output on Snapdragon X Elite Copilot+ PC
By VRAM tier · typical GPUs and Macs at Q4_K_M
Similar models
1| Model | Score | VRAM | tok/s | Why it is similar |
|---|---|---|---|---|
| MiniCPM V 4.7 1B DSparkOpenBMB | — | 0.5 GB | ~1489 | same size |
Specifications
- Publisher
- rinna
- Parameters
- 204M
- Architecture
- GPTNeoXForCausalLM
- Context Length
- 2,048 tokens
- Vocabulary Size
- 44,416
- Release Date
- 2022-08-31
- License
- MIT
Frequently Asked Questions
How much VRAM does Japanese GPT Neox Small need?
Japanese GPT Neox Small requires 0.1 GB of VRAM at Q4_K_M, or 0.5 GB at BF16.
VRAM = Weights + KV Cache + Overhead
Weights = 204M × 4.8 bits ÷ 8 = 0.1 GB
Fit ratings and hardware model lists check this model with room for a 16K-token context, which needs a little more memory.
VRAM usage by quantization
What's the best quantization for Japanese GPT Neox Small?
For Japanese GPT Neox Small, Q4_K_M (0.1 GB) offers the best balance of quality and VRAM usage. Q5_K_M (0.2 GB) provides better quality if you have the VRAM. The smallest option is Q2_K at 0.1 GB.
VRAM requirement by quantization
★ Recommended — best balance of quality and VRAM usage.
Can I run Japanese GPT Neox Small on a Mac?
Yes — MacBook Air 13" M3 (8 GB) and 145 other Macs can run Japanese GPT Neox Small. Apple Silicon uses unified memory, so the model shares RAM with the system. At Q4_K_M you need at least 0.1 GB of usable unified memory (RAM minus macOS overhead).
Can I run Japanese GPT Neox Small locally?
Yes — Japanese GPT Neox Small can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.
How fast is Japanese GPT Neox Small?
At Q4_K_M, Japanese GPT Neox Small can reach ~36923 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~5040 tok/s. 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 ÷ 0.13 × 0.65 = ~40000 tok/s
Estimated speed at Q4_K_M (0.1 GB)
Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
What's the download size of Japanese GPT Neox Small?
At Q4_K_M, the download is about 0.12 GB. The full-precision BF16 version is 0.41 GB. The smallest option (Q2_K) is 0.09 GB.
Which GPUs can run Japanese GPT Neox Small?
52 consumer GPUs can run Japanese GPT Neox Small at Q4_K_M. Top options include AMD Radeon RX 6700 XT, AMD Radeon RX 6800, AMD Radeon RX 6800 XT. 52 GPUs have plenty of headroom for comfortable inference.
Which devices can run Japanese GPT Neox Small?
163 devices with unified memory can run Japanese GPT Neox Small at Q4_K_M (0.1 GB), including AMD Ryzen AI 9 HX 370 (Strix Point) Laptop, ASUS Ascent GX10, 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.