--- title: 28.9M-parameter LLM runs locally on ESP32-S3 at 9 tokens/s description: Developer Slava S. has managed to run a 28.9M-parameter LLM entirely on an ESP32-S3 microcontroller at ~9 tokens/second, without relying on any cloud backend.... canonical: https://daily.dev/posts/28-9m-parameter-llm-runs-locally-on-esp32-s3-at-9-tokens-s-xmajabusu twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: 28.9M-parameter LLM runs locally on ESP32-S3 at 9 tokens/s | daily.dev og:description: Developer Slava S. has managed to run a 28.9M-parameter LLM entirely on an ESP32-S3 microcontroller at ~9 tokens/second, without relying on any cloud backend.... og:url: https://daily.dev/posts/28-9m-parameter-llm-runs-locally-on-esp32-s3-at-9-tokens-s-xmajabusu og:image: https://api.daily.dev/og/posts/XMAjabusU.png og:image:alt: 28.9M-parameter LLM runs locally on ESP32-S3 at 9 tokens/s og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # 28.9M-parameter LLM runs locally on ESP32-S3 at 9 tokens/s **[CNX Software](https://daily.dev/sources/cnx-software)** · [@cnxsoft](https://daily.dev/cnxsoft) · 3 min read · 0 upvotes · 0 comments ## Summary Developer Slava S. has managed to run a 28.9M-parameter LLM entirely on an ESP32-S3 microcontroller at ~9 tokens/second, without relying on any cloud backend. The key insight is using Google's Per-Layer Embeddings technique (from Gemma 3n/4): since most parameters reside in a large embedding table (~25M rows), they can be stored in slow 16MB flash memory rather than fast SRAM. Only the ~450 bytes needed per token are loaded into the 512KB SRAM at inference time. The model (14.9MB at 4-bit quantization) was trained on TinyStories and can only generate short stories — it cannot answer questions or follow instructions. The project runs on an ~$8 ESP32-S3 board with 8MB PSRAM and 16MB flash, connected to a small I2C display. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Similar posts on daily.dev - [Someone squeezed a 28.9M LLM onto an ESP32-S3, and so can you](https://daily.dev/posts/someone-squeezed-a-28-9m-llm-onto-an-esp32-s3-and-so-can-you-a5f9lzpva) · XDA Developers · 2 upvotes · 0 comments - [Someone wired up seven ESP32s to create a ~0.4B LLM, and so can you](https://daily.dev/posts/someone-wired-up-seven-esp32s-to-create-a-0-4b-llm-and-so-can-you-xwbd0nns2) · XDA Developers · 0 upvotes · 0 comments - [Running local models on an M4 with 24GB memory](https://daily.dev/posts/running-local-models-on-an-m4-with-24gb-memory-iezpeho6j) · Hacker News · 4 upvotes · 0 comments - [Small Language Models: Why 2026 Became the Year Bigger Stopped Meaning Better](https://daily.dev/posts/small-language-models-why-2026-became-the-year-bigger-stopped-meaning-better-omagodtq0) · Medium · 4 upvotes · 0 comments --- Tags: [#llm](https://daily.dev/tags/llm), [#iot](https://daily.dev/tags/iot), [#embedded](https://daily.dev/tags/embedded), [#local-ai](https://daily.dev/tags/local-ai) [View this post on daily.dev](https://daily.dev/posts/28-9m-parameter-llm-runs-locally-on-esp32-s3-at-9-tokens-s-xmajabusu) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"28.9M-parameter LLM runs locally on ESP32-S3 at 9 tokens/s","url":"https://daily.dev/posts/28-9m-parameter-llm-runs-locally-on-esp32-s3-at-9-tokens-s-xmajabusu","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/28-9m-parameter-llm-runs-locally-on-esp32-s3-at-9-tokens-s-xmajabusu"},"datePublished":"2026-08-03T10:40:43.828Z","dateModified":"2026-08-04T11:00:21.680Z","description":"Developer Slava S. has managed to run a 28.9M-parameter LLM entirely on an ESP32-S3 microcontroller at ~9 tokens/second, without relying on any cloud backend....","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/431c5d2eb44c4cdfe215cd2268b91525?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/431c5d2eb44c4cdfe215cd2268b91525?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"CNX Software","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Person","name":"CNXSoft","url":"https://daily.dev/cnxsoft","image":"https://avatars.githubusercontent.com/u/1366138?v=4","interactionStatistic":{"@type":"InteractionCounter","interactionType":{"@type":"EndorseAction"},"userInteractionCount":520}},"commentCount":0,"discussionUrl":"https://daily.dev/posts/28-9m-parameter-llm-runs-locally-on-esp32-s3-at-9-tokens-s-xmajabusu","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"llm,iot,embedded,local-ai","timeRequired":"PT3M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"CNX Software","item":"https://daily.dev/sources/cnx-software"},{"@type":"ListItem","position":3,"name":"28.9M-parameter LLM runs locally on ESP32-S3 at 9 tokens/s"}]} ```