Smol Llama 101M GQA — Hardware Requirements & GPU Compatibility
Chat
Smol Llama 101M GQA is a 101M-parameter open language model from BEE-spoke-data in the Llama family. It supports a context window of up to 1,024 tokens. At Q4_K_M it needs about 0.37 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)Best on 24 GB (Q4_K_M · ~1771 tok/s)Macs from 8 GB (Q4_K_M · ~194 tok/s)
Apache 2.0
Reference GPU · RTX 4090 24 GB
Estimates, ±20% · how we estimate
Run Smol Llama 101M GQA
One recommended way. Every other method, OS and quant is in the step-by-step guide.
Recommended
vllm serve BEE-spoke-data/smol_llama-101M-GQA --max-model-len 1024vLLM repo with its own quantisation
Download: Weights on Hugging Face
Other ways
each opens step-by-stepRun in cloud
Fits on RTX 3060 12GB (11 GB headroom) · Q4_K_M
- Generation speed
- ~632 tok/s
- generation speed
- Cost per 1M output tokens
- $0.036
- per 1M output tokens
How Much VRAM Does Smol Llama 101M GQA Need?
Select a quantization level; the listed hardware adheres to it.
| Quantization | VRAM | +32K | On RTX 4090 (reference) | |
|---|---|---|---|---|
| 0.4 GB | — | 2-bit · K-quant — 2-bit quantization with K-quant improvements | Fits: — context · ~1820 tok/s | |
| 0.4 GB | — | 3-bit medium — 3-bit medium quantization | Fits: — context · ~1820 tok/s | |
| 0.4 GB | — | 4-bit medium — 4-bit medium quantization — most popular sweet spot | Fits: — context · ~1771 tok/s | |
| 0.4 GB | — | 5-bit medium — 5-bit medium quantization — good quality/size tradeoff | Fits: — context · ~1724 tok/s | |
| 0.4 GB | — | 6-bit, very good — 6-bit quantization, very good quality | Fits: — context · ~1638 tok/s | |
| 0.4 GB | — | 8-bit, near-lossless — 8-bit quantization, near-lossless | Fits: — context · ~1598 tok/s | |
| 0.5 GB | — | 16-bit · bfloat — Brain floating point 16 — preferred for training | Fits: — context · ~1260 tok/s |
Selected Q4_K_M
4-bit medium quantization0.4GBat 1K context
- Weights
- 0.1 GB
- KV cache
- 0.0 GB
- Overhead
- 0.3 GB
Fits RTX 4090 (reference) · 23.6 GB free · ~1771 tok/s
≥50 tok/s needs ≥28 GB/s · RTX 5090 ~3148, RTX 3090 Ti ~1771
Which Hardware Can Run Smol Llama 101M GQA?
Q4_K_M · 0.4 GBSmol Llama 101M GQA (Q4_K_M) requires 0.4 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 Smol Llama 101M GQA, including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB).
≤ 11 GB
≤ 11 GB tier →fits · 7.6 GB free · up to — ctx
16 GPUs · 1 Mac class · 2 device families
RTX 3080~1336Simulate output on RTX 3080RTX 2080 Ti~1082Simulate output on RTX 2080 TiRTX 3070 Ti~1069Simulate output on RTX 3070 TiM3*8 GB~194Simulate output on iMac 24" M3 (8 GB)Orin Nano~179Simulate output on NVIDIA Jetson Orin Nano 8GB (Super)iPhone~129Simulate output on iPhone 1720 GB
fits · 19.6 GB free · up to — ctx
1 GPU
RX 7900 XT~1297Simulate output on RX 7900 XT32–40 GB
fits · 31.6 GB free · up to — ctx
1 GPU · 3 Mac classes · 4 device families
RTX 5090~3148Simulate output on RTX 5090M5 Max36 GB~870Simulate output on Mac Studio M5 Max (36 GB)M532 GB~291Simulate output on MacBook Air 13" M5 (32 GB)Orin~360Simulate output on NVIDIA Jetson AGX Orin 32GBSnapdragon X Elite (X1E-80-100)~237Simulate output on Snapdragon X Elite Copilot+ PC
By VRAM tier · typical GPUs and Macs at Q4_K_M
Alternatives to Smol Llama 101M GQA
Llama family · 19 models →Similar models
3| Model | Score | VRAM | tok/s | Why it is similar |
|---|---|---|---|---|
| Supra2 100M InstructSupraLabs | — | 0.4 GB | ~1680 | same size |
| Pythia 70M DedupedEleutherAI | — | 0.1 GB | ~10920 | same size |
| Macbert4csc Base Chineseshibing624 | — | 0.2 GB | ~2849 | same size |
Specifications
- Publisher
- BEE-spoke-data
- Family
- Llama
- Parameters
- 101M
- Architecture
- LlamaForCausalLM
- Context Length
- 1,024 tokens
- Vocabulary Size
- 32,128
- Release Date
- 2023-10-26
- License
- Apache 2.0
Frequently Asked Questions
How much VRAM does Smol Llama 101M GQA need?
Smol Llama 101M GQA requires 0.4 GB of VRAM at Q4_K_M, or 0.5 GB at BF16.
VRAM = Weights + KV Cache + Overhead
Weights = 101M × 4.8 bits ÷ 8 = 0.1 GB
KV Cache + Overhead ≈ 0.3 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.
VRAM usage by quantization
What's the best quantization for Smol Llama 101M GQA?
For Smol Llama 101M GQA, Q4_K_M (0.4 GB) offers the best balance of quality and VRAM usage. Q5_K_M (0.4 GB) provides better quality if you have the VRAM. The smallest option is Q2_K at 0.4 GB.
VRAM requirement by quantization
★ Recommended — best balance of quality and VRAM usage.
Can I run Smol Llama 101M GQA on a Mac?
Yes — MacBook Air 13" M3 (8 GB) and 145 other Macs can run Smol Llama 101M GQA. Apple Silicon uses unified memory, so the model shares RAM with the system. At Q4_K_M you need at least 0.4 GB of usable unified memory (RAM minus macOS overhead).
Can I run Smol Llama 101M GQA locally?
Yes — Smol Llama 101M GQA can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.
How fast is Smol Llama 101M GQA?
At Q4_K_M, Smol Llama 101M GQA can reach ~12973 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~1771 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 ÷ 0.37 × 0.65 = ~14054 tok/s
Estimated speed at Q4_K_M (0.4 GB)
Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
What's the download size of Smol Llama 101M GQA?
At Q4_K_M, the download is about 0.06 GB. The full-precision BF16 version is 0.20 GB. The smallest option (Q2_K) is 0.04 GB.
Which GPUs can run Smol Llama 101M GQA?
52 consumer GPUs can run Smol Llama 101M GQA 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 Smol Llama 101M GQA?
163 devices with unified memory can run Smol Llama 101M GQA at Q4_K_M (0.4 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.