The Only Rule That Matters
The model must fit entirely in RAM. If it doesn't, Ollama spills to disk and inference becomes unusably slow — we're talking minutes per response instead of seconds.
A rough sizing formula: model disk size in GB ≈ RAM needed. Add 2 GB headroom for the OS.
| VPS RAM | Max Model Size | Safe Choice |
|---|---|---|
| 4 GB | ~2 GB model | Llama 3.2 3B (Q4) |
| 8 GB | ~5 GB model | Llama 3.3 8B (Q4) |
| 16 GB | ~9 GB model | Qwen 2.5 14B (Q4) |
| 32 GB | ~20 GB model | DeepSeek R1 32B (Q4) |
Recommended Models by Use Case
General Chat and Summarization — Llama 3.3 8B
ollama pull llama3.1:8bThe most widely recommended starting model in 2026. Handles conversation, summarization, Q&A, and basic coding well. Requires 8 GB RAM. Inference on a 4-core CPU: 5–10 tokens/second.
Coding Assistant — Qwen 2.5 Coder 7B
ollama pull qwen2.5-coder:7bFine-tuned specifically for code. Outperforms Llama 3.3 on coding tasks at the same size. Requires 8 GB RAM.
Lightweight and Fast — Llama 3.2 3B
ollama pull llama3.2:3bFits in 4 GB RAM. Inference at 15–25 tokens/second on CPU — fast enough for interactive use. Quality is noticeably lower than 7B models but fine for simple tasks.
Reasoning and Math — Phi-4 14B
ollama pull phi4:14bMicrosoft's Phi-4 punches above its weight on structured reasoning and math. Regularly outperforms larger 30B models on benchmarks. Requires 16 GB RAM. Not suitable for long documents — 16K context window.
Privacy-Focused Tasks — Mistral 7B
ollama pull mistral:7bStrong general performance, small footprint. A reliable alternative to Llama 3.3 for teams that prefer European-origin models.
Quantization: Q4 vs Q8
Ollama defaults to Q4_K_M quantization when you pull a model. This cuts memory by 40–50% compared to full precision with minimal quality loss — roughly the difference between a RAW photo and a high-quality JPEG.
For most use cases, Q4_K_M is the right choice. If you need higher quality and have extra RAM, pull with :q8_0:
ollama pull llama3.1:8b-instruct-q8_0What CPU Inference Is Good For
On a 4-core CPU, expect 5–15 tokens per second for a 7B model. That's roughly one word per second at the low end. This is:
- Fine for background processing — summarizing documents, extracting data, classifying text
- Usable for interactive chat if you're patient
- Not suitable for real-time code completion or latency-sensitive applications
For real-time interactive use, 8+ cores and 16+ GB RAM gives noticeably better results.
Running these on NoctHost
The plan follows the model you pick from the table above: the Micro plan (1 vCPU, 1 GB) runs 1-3B models, the Standard plan (2 vCPU, 4 GB) is the sweet spot for 7B, and the Beast plan (6 vCPU, 16 GB) handles 13B and larger. Billing is hourly from a prepaid balance you top up with crypto, no card and no KYC, so you can size up for a heavy model and scale back down without a monthly commitment.