---
format: "aidr-story-markdown/v1"
id: "614a0336a7737e51afb3b077cdb47d041cfcc6e0785f6808a08729d3ded841df"
canonical_url: "https://aidr.today/614a0336?lang=en"
title: "Mercury 3: Diffusion at scale"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-10-19T17:27:11.000Z"
category: "Models"
topics: ["llm"]
source_urls: ["https://www.inceptionlabs.ai/blog/introducing-mercury-3","https://news.ycombinator.com/item?id=50155245"]
summary: "Mercury 3 is a diffusion LLM that outperforms GPT-6 Luna and Claude Haiku 5.5 while running 4–8x faster at 1,000+ tokens per second."
---

# Mercury 3: Diffusion at scale

> [Open the canonical story](<https://aidr.today/614a0336?lang=en>)

**Published:** 2026-10-19T17:27:11.000Z
**Category:** Models
**Topics:** llm

## Summary

Mercury 3 is a diffusion LLM that outperforms GPT\-6 Luna and Claude Haiku 5\.5 while running 4–8x faster at 1,000\+ tokens per second\.

## Sources

- [Story source](<https://www.inceptionlabs.ai/blog/introducing-mercury-3>)
- [Discussion](<https://news.ycombinator.com/item?id=50155245>)