01What Qwen 3.8 Max Is and How It Works

Qwen 3.8 Max is a powerful AI model with 2.4 trillion parameters and a 1 million context window. It is designed for agentic tasks, meaning it can work independently to complete complex goals. In this episode, Julian Goldie shows how to use Qwen 3.8 Max with Hermes Agent, which acts as the hands while Qwen provides the brain.

This model is comparable to Fable 5 in capability but has some advantages. It beats Fable 5 on key benchmarks and can be plugged directly into Hermes without needing an API, which you cannot do with Fable 5. Hermes Agent uses Qwen 3.8 Max to think, plan, and create, making it a versatile tool for coding, content creation, and more.

02Main Features of Qwen 3.8 Max

Qwen 3.8 Max comes with several standout features that make it useful for both beginners and advanced users. These features are highlighted in the episode:

  • 2.4 trillion parameters for deep reasoning and understanding.
  • 1 million context window, allowing it to handle large amounts of information.
  • Designed specifically for agentic tasks, meaning it can work autonomously.
  • Very fast response times, making it efficient for real-time use.
  • Strong coding abilities, tested with autonomous development over 10 days.
  • Compatible with Hermes Agent, including profiles, goal mode, and memory systems.
  • Works with MCPs (like Higgs Field MCP) to build apps and tools.

These features allow users to create games, apps, websites, videos, and more with minimal effort.

03How to Set Up Qwen 3.8 Max in Hermes Agent

  1. Open Hermes Agent and go to the manage tab or your dashboard.
  2. Click the update button to make sure you have the latest version (Hermes 0.2).
  3. Create a new profile specifically for Qwen 3.8 Max.
  4. Select Qwen 3.8 Max as the main model in the profile settings.
  5. Optionally, choose Qwen 3.8 Max through Open Router as an alternative.
  6. Save the profile, and you are ready to use Qwen 3.8 Max in Hermes.

This setup is simple and takes only a few minutes. Once set up, you can switch between different models like DeepSeek or Cloud Code whenever you want, using the same profile system.

04How to Use Qwen 3.8 Max for Coding and MCPs

In the episode, Julian demonstrates how to use Qwen 3.8 Max with Hermes Agent for various tasks. First, he shows how to connect it to an MCP like Higgs Field MCP. By giving the model an example image, it can generate code and build projects automatically.

This makes it easy to create functional apps without deep technical knowledge.

Another example is using the forward slash learn command. You can provide a guide, and Qwen 3.8 Max will learn the skill and save it as a skill file. This file is stored locally and can be reused later, saving time on repetitive tasks.

Additionally, you can use goal mode to set a target, and the model will work autonomously for hours to complete it.

For creativity, you can combine Qwen 3.8 Max with skills like Remotion to generate animated videos. Just give a simple prompt, and it produces a professional-looking video with b-roll, script, and editing done automatically.

05Results and Tests with Qwen 3.8 Max

Julian shares several test results in the episode. He tested 50 different builds using Qwen 3.8 Max, and it successfully created games, apps, tools, and websites. It also handled MCP integration smoothly, producing quality outputs.

The model ran autonomously for 10 days without handholding, evolving its own code from an empty folder to production-ready software. This shows its reliability for long-term projects. In terms of speed, responses are fast, making it practical for interactive sessions.

Compared to Fable 5, Qwen 3.8 Max reportedly beats it on some key benchmarks, though specific numbers are not provided. The episode also mentions combining Qwen 3.8 Max with other models like Fable 5 in Hermes mixture of agents to get even better results.

06Tips, Limits, and Mistakes to Avoid

To get the most out of Qwen 3.8 Max, Julian recommends using a memory system like Obsidian within Hermes. This helps the model maintain context across sessions and improves its performance, especially when switching between different models.

  • Always update Hermes to the latest version to ensure compatibility and speed.
  • Use the skill learning feature to train the model on specific guides for better outputs.
  • Take advantage of the workspace to save and review your projects, so you never lose your work.
  • Be aware that while Qwen 3.8 Max is powerful, complex tasks may still require human oversight.

One limitation mentioned is that the model is not yet perfect in all scenarios, but it is continuously improving. Avoid skipping the setup steps, as a proper configuration ensures smoother operation.

07Who Qwen 3.8 Max Is For and Why It Matters

Qwen 3.8 Max is for anyone who wants to leverage AI for coding, content creation, or automation. Whether you are a seasoned developer or a beginner, its user-friendly integration with Hermes Agent makes it accessible. In the episode, Julian emphasizes that it is designed for agentic tasks, meaning you can delegate complex jobs and let the AI handle them independently.

This matters because it saves time and effort. You can create apps, websites, and videos with just a few prompts, and even run autonomous projects while you focus on other things. The ability to switch between models and use memory systems adds flexibility, making it a valuable tool for both personal and professional use.

Ultimately, Qwen 3.8 Max represents a step forward in AI accessibility, allowing more people to build and innovate without needing advanced technical skills.