NEW DeepSeek V3.1 Update is WILD (FREE!) 🤯

Julian Goldie SEOAbout 7 min readMar 26, 2025Watch original
THE SUMMARYAI-generated

Deep Seek V3 0324 Update Summary

Key Concepts:

  • Deep Seek V3 0324: A new version of the Deep Seek AI model, positioned as a strong non-reasoning model.
  • Open Router: A platform offering free API access to Deep Seek V3 0324.
  • 3js: A JavaScript library used for creating 3D graphics in web browsers.
  • AI Agents: Automated systems powered by AI, capable of performing tasks autonomously.
  • N8N & Make.com: Workflow automation platforms used to create AI-powered automations.
  • Claude 3.7 Sonnet: A competing AI model used for comparison.
  • Reasoning Model: An AI model with the ability to think and reason.
  • Base Model: An AI model without reasoning capabilities.
  • API Key: A code used to authenticate and authorize access to an API.
  • Visual Studio Code: A popular code editor.
  • Client/Rode Code: Extensions for Visual Studio Code that allow users to interact with AI models.
  • Hugging Face: A platform for sharing and discovering machine learning models.

Deep Seek V3 0324 Overview

The video discusses the new Deep Seek V3 0324 (also referred to as Deep Seek 3.1) update, highlighting its improved performance compared to the previous version. It emphasizes the model's capabilities in coding, content creation, and reasoning, while also providing instructions on how to access it for free.

Accessing Deep Seek V3 0324

  • Deep Seek Website: Access the model directly through the Deep Seek website (deepseek.com). However, the server may be busy due to high demand.
  • LM Arena: Use Deep Seek V3 0324 via the LM Arena direct chat interface (lmarena.ai).
  • Open Router: Obtain a free API key from Open Router (openrouter.ai) and use the model in various applications. This is the preferred method due to its flexibility.
  • Hugging Face: Access the model through the Hugging Face platform.

Coding Capabilities

  • SAS Tool Creation: The video demonstrates the creation of a keyword research SAS tool in HTML using Deep Seek V3 0324. The model generated functional code quickly and efficiently.
    • Example: The model created a tool named "Keyword Genius" with features like keyword generation, competition scores, and cost-per-click data.
  • 3js Game Development: Deep Seek V3 0324 was used to create a 3D missile game with a single prompt. This showcases the model's ability to generate complex code with minimal input.
    • Example: The model created a game where the user collects rings in a 3D environment.
  • Website Design: The model was able to create a landing page for an SEO agency with a modern design.
    • Example: The landing page included a gradient background, information about the agency, and testimonials.

Content Creation

  • SEO Article Generation: Deep Seek V3 0324 was used to generate an SEO-optimized article with a keyword front-loaded title and a call to action.
    • The generated article was 485 words long and was rated as 0% AI detectable by ZeroGPT.
  • SEO Cost Calculator: The model created a functional SEO cost calculator with dropdown menus and cost calculations.

Reasoning Abilities

  • River Crossing Riddle: Deep Seek V3 0324 was tested with a reasoning challenge involving picking apples from a tree across a river in winter. The model provided a detailed analysis of the situation, considering factors like apple availability and potential solutions.
    • The model recognized that apples are not typically available in winter and offered alternative solutions.

Comparison with Claude 3.7 Sonnet

The video compares Deep Seek V3 0324 with Claude 3.7 Sonnet in terms of writing ability, reasoning, and coding.

  • Writing: Deep Seek V3 0324 produced more humanized content with less AI "fluff." Both models scored 0% AI detectable on ZeroGPT.
  • Reasoning: Deep Seek V3 0324 provided a more in-depth analysis of the river crossing riddle compared to Claude 3.7 Sonnet.
  • Coding: Deep Seek V3 0324 created a more complete and visually appealing landing page compared to Claude 3.7 Sonnet. However, Claude 3.7 Sonnet was better at recognizing the user's intent in certain coding tasks.
  • Web Search: Both models failed the web search test.
  • Pricing: Deep Seek V3 0324 is free to use, while Claude 3.7 Sonnet requires a paid subscription for full access.

Automating with Deep Seek V3 0324

The video demonstrates how to automate tasks using Deep Seek V3 0324 with various tools:

  • Visual Studio Code: Use the Client or Rode Code extensions to integrate Deep Seek V3 0324 into the code editor.
  • N8N: Create AI agents and automate workflows using the N8N platform.
    • Example: An AI agent was created to write an SEO article and send it to WordPress.
  • Make.com: Automate tasks using the Make.com platform.
    • Example: A workflow was created to generate content from a Google Sheet and post it to WordPress.
  • Browser Automation: Use the Browserless Web UI to control a web browser with Deep Seek V3 0324.

Conclusion

Deep Seek V3 0324 is a powerful and versatile AI model that offers a strong alternative to paid models like Claude 3.7 Sonnet. Its coding, content creation, and reasoning abilities, combined with its free accessibility, make it a valuable tool for developers, content creators, and anyone looking to automate tasks. While it may have some limitations, such as occasional server issues and slower response times, its overall performance and free access make it a compelling option.

Quen 2.5 VL 32b Model Summary

Key Concepts:

  • Quen 2.5 VL 32b: A new AI model from Alibaba, positioned as smarter and lighter than previous versions.
  • VL Series: A series of models more aligned with human preferences.
  • Open Router: A platform offering free API access to Quen 2.5 VL 32b.
  • N8N: A workflow automation platform used to create AI-powered automations.
  • Visual Studio Code: A popular code editor.
  • Client/Rode Code: Extensions for Visual Studio Code that allow users to interact with AI models.
  • Hugging Face: A platform for sharing and discovering machine learning models.
  • Video Generation: The ability of the model to generate videos from text prompts.
  • Thinking Mode: A feature to improve the prompt.
  • Artifacts: A feature to take the code and preview it once it's completed.

Quen 2.5 VL 32b Overview

The video discusses the new Quen 2.5 VL 32b model, highlighting its improved performance compared to the previous versions. It emphasizes the model's capabilities in coding, content creation, reasoning, and video generation, while also providing instructions on how to access it for free.

Accessing Quen 2.5 VL 32b

  • Quen Website: Access the model directly through the Quen website (chat.quen).
  • Open Router: Obtain a free API key from Open Router (openrouter.ai) and use the model in various applications.
  • Hugging Face: Access the model through the Hugging Face platform.

Key Features

  • Smarter and Lighter: The model is designed to be more efficient and aligned with human preferences.
  • Mathematics and Reasoning: The model excels in mathematical tasks and reasoning.
  • Video Generation: The model can generate videos from text prompts.
  • Search Feature: The model has a working search feature, allowing it to access real-time information.
  • Artifacts: The model has artifacts, allowing users to take the code and preview it once it's completed.

Coding Capabilities

  • Website Design: The model was able to create a landing page for an SEO agency.
  • SEO Tool Creation: The model was able to create a basic SEO tool.

Content Creation

  • SEO Article Generation: The model was used to generate an SEO-optimized article.

Comparison with Other Models

The video compares Quen 2.5 VL 32b with other models like M out 3.1 and Google's Gemma 3, highlighting its superior performance in various benchmarks.

Automating with Quen 2.5 VL 32b

The video demonstrates how to automate tasks using Quen 2.5 VL 32b with various tools:

  • Visual Studio Code: Use the Client or Rode Code extensions to integrate Quen 2.5 VL 32b into the code editor.
  • N8N: Create AI agents and automate workflows using the N8N platform.
    • Example: An AI agent was created to write an SEO article and send it to WordPress.

Conclusion

Quen 2.5 VL 32b is a powerful and versatile AI model that offers a strong alternative to paid models. Its coding, content creation, reasoning, and video generation abilities, combined with its free accessibility, make it a valuable tool for developers, content creators, and anyone looking to automate tasks.

AI summaries can miss context or contain errors. Check important details against the original video.

Go a little deeper.

Have a question about this video? Load its transcript to open the video chat.