@xiaohu: Yesterday, I saw many people sharing Apodex 1.1, an AI agent specifically built for deep research to solve those hard problems that 'have no ready-made answers and require extensive investigation'. Curious, I tested it with two tasks, and they ran all afternoon without finishing. The execution time is indeed long. This agent can, as long as you give it a goal, run for an extended period…

X AI KOLs Timeline News
ai-agent deep-research automation complex-tasks sub-agent-system task-decomposition

Summary

Apodex 1.1 is an AI agent designed for deep research, capable of handling complex tasks that require extensive investigation. It uses a main agent to decompose problems and asynchronously dispatches multiple sub-agents for execution, supporting long-running operations and automatic recovery.

Yesterday, I saw many people sharing Apodex 1.1 An AI agent specifically built for deep research Solving those hard problems that 'have no ready-made answers and require extensive investigation' Curious, I tested it with two tasks, and they ran all afternoon without finishing The execution time is really long This agent can, as long as you give it a goal, run for a long time, automatically recover from failures, and verify if the delivered results are accurate After receiving a task, the main agent decomposes it into various sub-issues and asynchronously dispatches specialized sub-agents for execution. Each sub-agent has its own independent context, prompts, and toolset. The reports from sub-agents are compiled into a shared report pool, and the orchestrator reads asynchronously without being blocked by the slowest one. A single task can schedule up to 150 sub-agents What problem does it solve? Past: One question, one answer, one report. Evaluation criteria were usually: Is the answer correct? Is knowledge coverage sufficient? Are there many citations? Is the report well-written and complete? New requirements Apodex aims to propose: A complete task from input to delivery, with evaluation criteria changed to: Does it understand the goal? Can it operate on real files? Can it invoke code and tools? Can it maintain long-task states? Can it adjust locally when changes occur? Can it recover automatically from execution failures? Is the final conclusion verifiable? Therefore, it truly challenges the current product form of Deep Research: Collecting information and generating reports only covers part of complex tasks. Real professional tasks also require reading files, cleaning data, selecting methods, executing code, handling exceptions, and verifying conclusions. So it's especially suitable for: researchers, analysts, professional users
Original Article
View Cached Full Text

Cached at: 08/27/26, 01:28 AM

Yesterday I saw many people sharing Apodex 1.1

An agent specifically designed for deep research

It targets those tough problems that “have no ready-made answers and require extensive investigation to solve”

Curious, I tested it and ran two tasks—all afternoon without completion

The execution time is really long

What this tool does is you give it an objective, and it can run for an extended period, automatically recover from failures, and verify the accuracy of its deliverables

Upon receiving a task, the main agent breaks it down into various sub-problems and dispatches them asynchronously to specialized sub-agents for execution. Each sub-agent has its own independent context, prompts, and toolset

The sub-agents’ reports are fed into a shared reporting pool. The orchestrator reads asynchronously and isn’t held back by the slowest one. A single task can coordinate up to 150 sub-agents

What problem does this solve?

Previously:

One query, one response, one report. Success was typically measured by: Accuracy of the answer Breadth of knowledge covered Number of citations Completeness of the report

Apodex proposes new requirements:

A complete task from input to deliverable, where success is measured by:

Understanding the goal Ability to operate real files Ability to invoke code and tools Ability to maintain long-task state Ability to adapt locally when changes occur Self-recovery upon execution failure Whether the final conclusions can be verified

Thus, it truly challenges the current product form of Deep Research: Collecting materials and generating reports only covers part of a complex task

Genuine professional tasks also require reading files, cleaning data, selecting methods, executing code, handling exceptions, and verifying conclusions

So it’s especially suited for: researchers, analysts, and professional users

Similar Articles

@Apodex_AI: Dive in Blog: https://apodex.com/blog/apodex-1.0 Tech report: http://apodex.com/pdf/20260608 Github: https://github.com…

X AI KOLs Following

ApodexAI releases Apodex-1.0, a deep-research model that operates as a tool-using ReAct agent. Its heavy-duty mode, Apodex-1.0-H, uses an asynchronous agent team with up to 150 sub-agents and achieves new state-of-the-art results on deep-research benchmarks including BrowseComp, DeepSearchQA, HLE, and FrontierScience, surpassing models like GPT-5.5-pro and Claude-Opus-4.8.

@Xudong07452910: Recently, while looking into agent self-evolution research, I've been focusing on a question: When are the trajectories left by agents worth continued learning for the next round of training? This time, I took a self-evolving agent paper from SEED and ran it entirely in Apodex…

X AI KOLs Timeline

This article discusses the importance of trajectory learning in AI agent self-evolution and introduces the Apodex 1.1 system and open-source tool FrontierAgent for executing and evaluating long-duration research tasks.

@justloveabit: With This Open-Source Tool, I Got a Team of AIs to Work for Me. Here's the deal: I've been tinkering with various AI agents lately. Multiple Claude Code windows open, Codex running, occasionally using Cursor. The result? Total chaos—I had no idea what each agent was doing or how much it was costing. Restar…

X AI KOLs Timeline

This article introduces Paperclip, an open-source tool designed to centrally manage and orchestrate multiple AI agents. By simulating a corporate organizational structure, task assignment, and budget control, it addresses key pain points in multi-agent collaboration, such as lost context, unpredictable costs, and chaotic scheduling.

@sunmer575399: Everyone doing AI should check out hermes-agent. Not because of the 204.0k stars. But because it solves the core problems cleanly. It's known as the 'agent that grows' — it doesn't stick to a fixed workflow, but automatically adjusts strategies based on your usage habits... I've been running it for two weeks, and it feels like it understands you more the more you use it, much more flexible than those rigid auto...

X AI KOLs Timeline

Hermes Agent is a self-improving AI agent framework developed by Nous Research, supporting multi-environment deployment and model switching, aimed at improving development efficiency and user experience.