---
format: "aidr-story-markdown/v1"
id: "32e2871abf0fad059f6ea69a6936d1a045f458da582e32614392cb434a26c28f"
canonical_url: "https://aidr.today/32e2871a?lang=en"
title: "Marin 535B Model Hits 7% Training Mark Ahead of Sept 1 Review"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-08-28T05:47:07.000Z"
category: "Models"
topics: ["llm","training","research"]
source_urls: ["https://huggingnews.com/ai/marin-535b-model-hits-7percent-training-mark-ahead-of-sept-1-review-9e3e4b93","https://x.com/percyliang/status/2093139094968922440","https://x.com/stanfordnlp/status/2093165946965987338"]
summary: "The Marin 535B-A23B model has reached an early milestone in its development process with training currently on track. Percy Liang stated the run is roughly 7% finished, confirming the large-scale AI project is proceeding as planned. The Marin team will host a Zoom panel on September 1 at 10 PT to discuss the design decisions, tradeoffs, and learnings from the training run. Details for the session are available through the project's Discord server."
---

# Marin 535B Model Hits 7% Training Mark Ahead of Sept 1 Review

> [Open the canonical story](<https://aidr.today/32e2871a?lang=en>)

**Published:** 2026-08-28T05:47:07.000Z
**Category:** Models
**Topics:** llm, training, research

## Summary

The Marin 535B\-A23B model has reached an early milestone in its development process with training currently on track\. Percy Liang stated the run is roughly 7% finished, confirming the large\-scale AI project is proceeding as planned\. The Marin team will host a Zoom panel on September 1 at 10 PT to discuss the design decisions, tradeoffs, and learnings from the training run\. Details for the session are available through the project's Discord server\.

## Sources

- [Story source](<https://huggingnews.com/ai/marin-535b-model-hits-7percent-training-mark-ahead-of-sept-1-review-9e3e4b93>)
- [Story source](<https://x.com/percyliang/status/2093139094968922440>)
- [Supporting source](<https://x.com/stanfordnlp/status/2093165946965987338>)

