Developer runs a 28.9M-parameter LLM on an $8 ESP32-S3 chip offline
joemeno · x · 2026-07-25
A developer has forced a 28.9-million-parameter language model onto a standard ESP32-S3 microcontroller that costs about $8.
- The model runs fully offline and produces coherent short stories at around 9.5 tokens per second.
- The chip has only 512 KB of fast SRAM and 16 MB of flash, so the key trick is architectural: most of the embedding table was moved into flash and memory-mapped.
- The runtime only needs to pull about 450 bytes per token, keeping active working memory in SRAM.
- The result is a capable local language model on hardware that draws roughly the power of a single LED.
- The thread also points to early physical-device experiments, including a minimalist voice-controlled universal remote.
Related event: $8 ESP32-S3 Microcontroller Runs LLM Offline(4 posts)→
More from coding & agent
- Claude Code 2.1.220 adds 6,256 prompt tokens and shifts system-token share higher — ClaudeCodeLog · 2026-07-25
- Claude Code CLI 2.1.220 ships with crash fixes and stability improvements — ClaudeCodeLog · 2026-07-25
- Grok Build update adds guided onboarding, search overrides, and failed-run resume — elonmusk · 2026-07-25
- Anthropic releases Claude Code v2.1.220 with bug fixes and reliability improvements — ashwin-ant · 2026-07-25
- A Mac coding-agent benchmark says the harness can cut token use 5× — asankhs · 2026-07-25
- Andrew Ng releases a free one-hour course on building agentic knowledge graphs — goyalshaliniuk · 2026-07-25