---
title: "Discussion on \" I Benchmarked 3 Local LLMs on My Laptop — Here's What the Numbers Actually Show\" | Hashnode"
description: "Discussion on \" I Benchmarked 3 Local LLMs on My Laptop — Here's What the Numbers Actually Show\". Run llama3.2, phi3, mistral locally with Ollama. Benchmark P99 latency, tokens/sec, and memory. FastAPI wrapper with JSON validation included."
url: "https://hashnode.com/posts/i-benchmarked-3-local-llms-on-my-laptop-here-s-what-the-numbers-actually-show/6a22f263857478689f4ff109"
---

[Feed](https://hashnode.com/feed) /

Discussion

# [I Benchmarked 3 Local LLMs on My Laptop — Here's What the Numbers Actually Show](https://thinkwithops.hashnode.dev/i-benchmarked-3-local-llms-on-my-laptop-here-s-what-the-numbers-actually-show?utm_source=hashnode&utm_medium=feed)

Run llama3.2, phi3, mistral locally with Ollama. Benchmark P99 latency, tokens/sec, and memory. FastAPI wrapper with JSON validation included.

Jun 5 · 6 min read · thinkwithops.hashnode.dev  
Read the article

-   [#ai](https://hashnode.com/tag/ai)
-   [#python](https://hashnode.com/tag/python)
-   [#machine-learning](https://hashnode.com/tag/machine-learning)
-   [#ollama](https://hashnode.com/tag/ollama)

## Written by

[Vijaya Rajeev Bollu](https://hashnode.com/@thinkwithops)

@thinkwithops · Kitchener,Canada

  
[View profile](https://hashnode.com/@thinkwithops)

I explain how software and AI systems really work, then build them: DevOps, SRE, ML engineering, MLOps and LLMOps. ThinkWithOps on YouTube.

## Responses

Markdown supported · ⌘↵ to post

No responses yet.
