---
format: "aidr-story-markdown/v1"
id: "4b4617fac923eecfd4e8bfc59cdb52d60135128b3d44a09b79f648b019bd84ff"
canonical_url: "https://aidr.today/4b4617fa?lang=en"
title: "Nanbeige Releases 33B Parameter Model With State-of-the-Art Agentic Performance"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-10-25T12:38:15.000Z"
category: "Models"
topics: ["agent","llm","open-source"]
source_urls: ["https://marketbrief.now/ai/nanbeige-releases-33b-parameter-model-with-state-of-the-art-agentic-perf-49ffccaa","https://huggingface.co/Nanbeige/Nanbeige4.3-33B","https://x.com/nanbeige/status/2114328933181333798","https://x.com/teortaxesTex/status/2114329947913535770","https://huggingnews.com/ai/nanbeige-releases-33b-parameter-model-with-state-of-the-art-agentic-perf-49ffccaa","https://www.marktechpost.com/2026/10/25/nanbeige-releases-nanbeige4-3-33b-open-agentic-coding-model/"]
summary: "Nanbeige has released the Nanbeige4.3-33B, an open dense model with approximately 30B parameters designed for state-of-the-art agentic performance among open models, with strengths in agentic coding and cowork benchmarks. The model will serve as a teacher to train smaller versions, starting with the Nanbeige4.3-3B, and reportedly handles real-world tasks using cost-efficient inference when integrated into mainstream agent harnesses."
---

# Nanbeige Releases 33B Parameter Model With State\-of\-the\-Art Agentic Performance

> [Open the canonical story](<https://aidr.today/4b4617fa?lang=en>)

**Published:** 2026-10-25T12:38:15.000Z
**Category:** Models
**Topics:** agent, llm, open\-source

## Summary

Nanbeige has released the Nanbeige4\.3\-33B, an open dense model with approximately 30B parameters designed for state\-of\-the\-art agentic performance among open models, with strengths in agentic coding and cowork benchmarks\. The model will serve as a teacher to train smaller versions, starting with the Nanbeige4\.3\-3B, and reportedly handles real\-world tasks using cost\-efficient inference when integrated into mainstream agent harnesses\.

## Sources

- [Story source](<https://marketbrief.now/ai/nanbeige-releases-33b-parameter-model-with-state-of-the-art-agentic-perf-49ffccaa>)
- [Story source](<https://huggingface.co/Nanbeige/Nanbeige4.3-33B>)
- [Story source](<https://x.com/nanbeige/status/2114328933181333798>)
- [Supporting source](<https://x.com/teortaxesTex/status/2114329947913535770>)
- [Story source](<https://huggingnews.com/ai/nanbeige-releases-33b-parameter-model-with-state-of-the-art-agentic-perf-49ffccaa>)
- [Story source](<https://www.marktechpost.com/2026/10/25/nanbeige-releases-nanbeige4-3-33b-open-agentic-coding-model/>)