Sarashina2.2 0.5B Instruct v0.1 — Hardware Requirements & GPU Compatibility
ChatFunctions
Sarashina2.2 0.5B Instruct v0.1 is a 793M-parameter open language model from sbintuitions. It supports a context window of up to 8,192 tokens. At Q4_K_M it needs about 0.90 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 · 8K)Best on 24 GB (Q4_K_M · ~728 tok/s · 8K)Macs from 8 GB (Q4_K_M · ~80 tok/s)
MIT
Reference GPU · RTX 4090 24 GB
Estimates, ±20% · how we estimate
Run Sarashina2.2 0.5B Instruct v0.1
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
ollama run hf.co/mmnga/sarashina2.2-0.5b-instruct-v0.1-gguf:Q4_K_MQ4_K_M fits fully on the RTX 4090 · 22.7 GB free · 8K context
Download: sarashina2.2-0.5b-instruct-v0.1-Q4_K_M.gguf0.53 GBmmnga
Ways to runOther ways
each opens step-by-stepRun in cloud
Your own private copy · Q4_K_M · 24 GB GPU
- The easiest way to try this model
- Running in a few minutes, easy setup
- Chat right here or connect any app
$5+ bonus for new users
Runs on RTX 3080 Ti (12 GB) (10 GB headroom) · Q4_K_M
- Generation speed
- ~659 tok/s
- generation speed
- Cost per 1M output tokens
- $0.052
- per 1M output tokens
How Much VRAM Does Sarashina2.2 0.5B Instruct v0.1 Need?
Select a quantization level; the listed hardware adheres to it.
| Quantization | VRAM | +32K | On RTX 4090 (reference) | |
|---|---|---|---|---|
| 0.8 GB | — | 2-bit · K-quant — 2-bit quantization with K-quant improvements | Fits: 8K context · ~862 tok/s | |
| 0.8 GB | — | 3-bit small — 3-bit small quantization | Fits: 8K context · ~851 tok/s | |
| 0.8 GB | — | 3-bit medium — 3-bit medium quantization | Fits: 8K context · ~809 tok/s | |
| 0.8 GB | — | 4-bit · legacy — 4-bit legacy quantization | Fits: 8K context · ~799 tok/s | |
| 0.9 GB | — | 4-bit medium — 4-bit medium quantization — most popular sweet spot | Fits: 8K context · ~728 tok/s | |
| 1.0 GB | — | 5-bit medium — 5-bit medium quantization — good quality/size tradeoff | Fits: 8K context · ~662 tok/s | |
| 1.1 GB | — | 6-bit, very good — 6-bit quantization, very good quality | Fits: 8K context · ~607 tok/s | |
| 1.2 GB | — | 8-bit, near-lossless — 8-bit quantization, near-lossless | Fits: 8K context · ~537 tok/s |
Selected Q4_K_M
4-bit medium quantization1.3GBat 8K context
- Weights
- 0.5 GB
- KV cache
- 0.5 GB
- Overhead
- 0.3 GB
Fits RTX 4090 (reference) · 22.7 GB free · ~728 tok/s
≥50 tok/s needs ≥69 GB/s · RTX 5090 ~1294, RTX 3090 Ti ~728
Which Hardware Can Run Sarashina2.2 0.5B Instruct v0.1?
Q4_K_M · 0.9 GBSarashina2.2 0.5B Instruct v0.1 (Q4_K_M) requires 0.9 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 2+ GB is recommended. Using the full 8K context window can add up to 0.4 GB, bringing total usage to 1.3 GB. 52 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti. 163 devices with unified memory can run Sarashina2.2 0.5B Instruct v0.1, including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB).
≤ 11 GB
≤ 11 GB tier →fits · 6.7 GB free · up to 8K ctx
16 GPUs · 1 Mac class · 2 device families
RTX 3080~549Simulate output on RTX 3080RTX 2080 Ti~445Simulate output on RTX 2080 TiRTX 3070 Ti~439Simulate output on RTX 3070 TiM3*8 GB~80Simulate output on iMac 24" M3 (8 GB)Orin Nano~74Simulate output on NVIDIA Jetson Orin Nano 8GB (Super)iPhone~53Simulate output on iPhone 1720 GB
fits · 18.7 GB free · up to 8K ctx
1 GPU
RX 7900 XT~533Simulate output on RX 7900 XT32–40 GB
fits · 30.7 GB free · up to 8K ctx
1 GPU · 3 Mac classes · 4 device families
RTX 5090~1294Simulate output on RTX 5090M5 Max36 GB~358Simulate output on Mac Studio M5 Max (36 GB)M532 GB~120Simulate output on MacBook Air 13" M5 (32 GB)Orin~148Simulate output on NVIDIA Jetson AGX Orin 32GBSnapdragon X Elite (X1E-80-100)~98Simulate 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 |
|---|---|---|---|---|
| Qwen3 0.6BAlibaba | — | 1.7 GB | ~753 | same size |
Specifications
- Publisher
- sbintuitions
- Parameters
- 793M
- Architecture
- LlamaForCausalLM
- Context Length
- 8,192 tokens
- Vocabulary Size
- 102,400
- Release Date
- 2025-02-26
- License
- MIT
Frequently Asked Questions
How much VRAM does Sarashina2.2 0.5B Instruct v0.1 need?
Sarashina2.2 0.5B Instruct v0.1 requires 0.9 GB of VRAM at Q4_K_M, or 2.0 GB at BF16.
VRAM = Weights + KV Cache + Overhead
Weights = 793M × 4.8 bits ÷ 8 = 0.5 GB
KV Cache + Overhead ≈ 0.4 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.
KV Cache + Overhead ≈ 0.8 GB (at full 8K context)
VRAM usage by quantization
What's the best quantization for Sarashina2.2 0.5B Instruct v0.1?
For Sarashina2.2 0.5B Instruct v0.1, Q4_K_M (0.9 GB) offers the best balance of quality and VRAM usage. Q5_0 (0.9 GB) provides better quality if you have the VRAM. The smallest option is IQ2_XXS at 0.6 GB.
VRAM requirement by quantization
★ Recommended — best balance of quality and VRAM usage.
Can I run Sarashina2.2 0.5B Instruct v0.1 on a Mac?
Yes — MacBook Air 13" M3 (8 GB) and 145 other Macs can run Sarashina2.2 0.5B Instruct v0.1. Apple Silicon uses unified memory, so the model shares RAM with the system. At Q4_K_M you need at least 0.9 GB of usable unified memory (RAM minus macOS overhead).
Can I run Sarashina2.2 0.5B Instruct v0.1 locally?
Yes — Sarashina2.2 0.5B Instruct v0.1 can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.
How fast is Sarashina2.2 0.5B Instruct v0.1?
At Q4_K_M, Sarashina2.2 0.5B Instruct v0.1 can reach ~5333 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~728 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 ÷ 0.9 × 0.65 = ~5778 tok/s
Estimated speed at Q4_K_M (0.9 GB)
Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
What's the download size of Sarashina2.2 0.5B Instruct v0.1?
At Q4_K_M, the download is about 0.48 GB. The full-precision BF16 version is 1.59 GB. The smallest option (IQ2_XXS) is 0.22 GB.
Which GPUs can run Sarashina2.2 0.5B Instruct v0.1?
52 consumer GPUs can run Sarashina2.2 0.5B Instruct v0.1 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 Sarashina2.2 0.5B Instruct v0.1?
163 devices with unified memory can run Sarashina2.2 0.5B Instruct v0.1 at Q4_K_M (0.9 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.