---
format: "aidr-story-markdown/v1"
id: "6161f7c8ceafd471f905039db8605e0488db6ce77c7f52d6744150e9df8aefd7"
canonical_url: "https://aidr.today/6161f7c8?lang=en"
title: "Apodex Launches 1.1 Model Family and Open Source Workbench for Professional AI Agents"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-08-25T08:45:32.000Z"
category: "Models"
topics: ["agent","multi-agent","open-source","inference"]
source_urls: ["https://huggingnews.com/ai/apodex-launches-11-model-family-and-open-source-workbench-for-profession-1d333cf6","https://x.com/Apodex_AI/status/2091916791308313018","https://x.com/sgl_project/status/2092143641980698932","https://x.com/sgl_project/status/2092145313461281217","https://x.com/sgl_project/status/2092144259503923642"]
summary: "The new model family from Apodex includes a frontier-level AI for deep search, a mini version for local deployment, and an open-source research workbench called FrontierAgent. The series coordinates multiple agents in parallel through an asynchronous team structure, allowing for the integration of findings from various agents while permitting users to redirect the work in real time. The models target scientific research, financial analysis, and deep search. SGLang has released deployment recipes for the mini variant, including an FP8 version with a 262,144 token context length and the use of radix cache to optimize asynchronous workflows."
---

# Apodex Launches 1\.1 Model Family and Open Source Workbench for Professional AI Agents

> [Open the canonical story](<https://aidr.today/6161f7c8?lang=en>)

**Published:** 2026-08-25T08:45:32.000Z
**Category:** Models
**Topics:** agent, multi\-agent, open\-source, inference

## Summary

The new model family from Apodex includes a frontier\-level AI for deep search, a mini version for local deployment, and an open\-source research workbench called FrontierAgent\. The series coordinates multiple agents in parallel through an asynchronous team structure, allowing for the integration of findings from various agents while permitting users to redirect the work in real time\. The models target scientific research, financial analysis, and deep search\. SGLang has released deployment recipes for the mini variant, including an FP8 version with a 262,144 token context length and the use of radix cache to optimize asynchronous workflows\.

## Sources

- [Story source](<https://huggingnews.com/ai/apodex-launches-11-model-family-and-open-source-workbench-for-profession-1d333cf6>)
- [Story source](<https://x.com/Apodex_AI/status/2091916791308313018>)
- [Supporting source](<https://x.com/sgl_project/status/2092143641980698932>)
- [Supporting source](<https://x.com/sgl_project/status/2092145313461281217>)
- [Supporting source](<https://x.com/sgl_project/status/2092144259503923642>)