AI Daily: Gemini 3.1 Pro Doubles Reasoning Power! Claude Enters PPT - A Summary of Highlights

Google Gemini 3.1 Pro Makes a Strong Debut: Doubling Reasoning Capabilities and Generating SVG Animations. What Else is New with Claude in PPT?

If you thought last week’s AI progress wasn’t fast enough, this week’s news will definitely make your head spin. Google just dropped a bombshell—Gemini 3.1 Pro is officially released, and this isn’t just a minor patch; it boasts a significant leap in logical reasoning and generative capabilities. At the same time, Anthropic isn’t falling behind, extending Claude’s reach into the office’s most commonly used PowerPoint, while also raising a red flag on developer API usage guidelines.

Are you ready to see how these new tools are changing workflows? Here is a summary of the most important AI developments this week.

Google Gemini 3.1 Pro: The “Brainpower” Born to Solve Complex Puzzles

Remember last week when Google released Gemini 3 Deep Think? That was mainly a warm-up for the scientific and research fields. Today, Google officially released the core intelligence of all these breakthroughs—Gemini 3.1 Pro. The design intent of this model is very clear: when simple answers are not enough to solve a problem, it’s time for it to step up.

A Massive Leap in Logical Reasoning Capabilities

To be honest, what people care most about with model upgrades is usually, “Did it get smarter?” According to the latest announcement on the Google official blog, Gemini 3.1 Pro has shown astonishing growth in its ability to handle completely new logical patterns.

In the ARC-AGI-2 benchmark, a rigorous test designed specifically to evaluate a model’s ability to solve “never-before-seen logic puzzles,” Gemini 3.1 Pro achieved a high score of 77.1%. What does this mean? Its reasoning performance is more than double that of its predecessor, the 3 Pro. For tasks that require multi-step thinking, comprehensive data analysis, or deep comprehension of long texts, this is undoubtedly a massive upgrade.

Text-to-SVG Animation: A New Toy for Designers

Besides getting smarter, it has also become “more creative.” There is a very cool feature in this update: code-based animation generation.

In the past, generating images meant generating pixels; zooming in would make it blurry, and the file size was large. But Gemini 3.1 Pro can now generate SVG format animations directly from text prompts. Because these animations are built purely from code, no matter how much you zoom in or out, the image will always remain crisp and sharp, and the file size is extremely small, making it perfect for web design. Imagine getting a dynamic icon that can be directly used on a website just by entering a description. This is simply a time-saving artifact for front-end developers.

Fully Open for Trials

The good news is that this model has already started rolling out to various platforms:

• Developers: Can be the first to experience it via Google AI Studio and the Gemini API.
• Enterprise Users: Vertex AI and Gemini Enterprise have been updated simultaneously.
• General Users: Open your Gemini App or NotebookLM now, and you can feel the power of this new assistant.

Claude Officially Enters PowerPoint: A Brand New Plug-in for Presentation Creation

For office workers struggling in a sea of slides every day, this is definitely exciting news. Anthropic announced that Claude in PowerPoint is now available to Pro plan users.

Linking Real Data via Connectors

This is not just a simple chatbot sidebar. The biggest highlight of this update is its support for Connectors. What does this mean? It means you don’t need to copy and paste back and forth between your browser and PPT.

Right within PowerPoint, you can have Claude extract context information from your daily tools (like project management software or document libraries) and directly transform this data into slide content. For project managers who need to frequently update business reports and integrate data from multiple sources, this can save a massive amount of trivial time. Although it’s currently exclusive to Pro users, it demonstrates a significant step for AI from “dialogue” to “execution.”

Attention Developers: The Red Line for Claude Code Authentication and Credential Usage

Alongside feature updates, Anthropic also issued an important announcement regarding security and compliance, especially for developers using Claude Code.

According to the latest Claude Code Legal and Compliance document, the official guidelines clearly distinguish between use cases for OAuth and API Keys, and draw a line that must not be crossed.

OAuth vs. API Key: Stop Mixing Them Up

Anthropic emphasizes that OAuth authentication (typically used for Free, Pro, and Max account logins) is specifically designed for official Claude Code tools and Claude.ai.

• Prohibited Actions: Developers must not use OAuth tokens obtained via personal accounts (Free/Pro/Max) in any third-party products, tools, or services (including when using the Agent SDK). This is considered a violation of the consumer Terms of Service.
• Correct Approach: If you are building products or services that interact with Claude, you must use an API Key obtained through the Claude Console or a supported cloud provider.

Simply put, officially, third-party developers are not allowed to “borrow” a user’s personal account permissions to run applications. This is both to protect the business model and, more importantly, to safeguard user data security. Anthropic has stated it reserves the right to take measures to enforce these restrictions, potentially without prior notice. So, if you are a developer, now is the time to check if your authentication methods are compliant.

Frequently Asked Questions (FAQ)

To help everyone quickly understand this update, we’ve compiled a few of the most critical questions:

1. Is Gemini 3.1 Pro currently free to use?

Gemini 3.1 Pro is currently rolled out in the Gemini App. Some features might require a subscription to the Google One AI Premium plan (i.e., using the Advanced version) for a full experience, especially those involving advanced reasoning or specific integrations. Developers can test it during the preview period via the API; please refer to Google AI Studio for specific pricing.

2. What’s the difference between generating SVG animations and generating images?

Traditional AI image generation creates JPG or PNG images composed of pixels, which distort when scaled up and cannot have their internal structures easily edited. What Gemini 3.1 Pro generates is an SVG (Scalable Vector Graphics), which is essentially a piece of code. This means you can freely scale the image without any loss of quality, and even directly modify the code to adjust animation details. The file size is also very small, making it extremely suitable for web development.

3. Can I use my Claude Pro account to develop my own automated tools?

This depends on how you connect. If you are using official tools provided by Claude (like the Claude Code CLI), authenticating with your Pro account is fine. However, if you are writing your own programs (e.g., using Python to call the Agent SDK), according to the new guidelines, you cannot use the OAuth token from your personal account. Instead, you should apply for and use an API Key. Although this may incur extra API costs, it is the only compliant path.

4. Do I need to install extra software for the Claude feature in PowerPoint?

Usually, these types of integrations are installed via the Microsoft Add-in store. You need an active Claude Pro subscription and must log into your account within PowerPoint to enable the related features. Please ensure your version of PowerPoint is up-to-date to ensure compatibility.