Current Capabilities

14 capabilities across today's leading AI platforms

AI capabilities have expanded dramatically. As of early 2026, the three leading AI platforms (Anthropic's Claude, OpenAI's ChatGPT, and Google's Gemini) share a common set of 14 core capabilities. Understanding what these tools can do today is essential for planning how your institution will use them.

The Three Frontier Models

Before diving into capabilities, it helps to know the major players:

ProviderPlatformFlagship ModelSpecialty Tools
AnthropicClaudeOpus 4.5Claude.ai, Claude Code
OpenAIChatGPTGPT-5.2 ProChatGPT, Codex
GoogleGemini3 ProGemini, NotebookLM

Each platform has unique strengths, but they all share the core capabilities listed below. Your choice of platform matters less than your ability to use these capabilities effectively.

14 Core Capabilities

1. Text Chat

The foundational capability. Ask questions, get answers, have conversations. Every platform supports multi-turn dialogue where the AI remembers context from earlier in the conversation.

Education example: Drafting communications, brainstorming solutions, getting explanations of complex regulations.

2. File and Document Creation

AI can generate complete documents: Word files, spreadsheets, PDFs, and more. You provide the requirements, and AI builds the document structure and content.

Education example: Creating board report templates, policy drafts, budget summaries, or meeting agendas.

3. Deep Research

Extended reasoning combined with web search allows AI to research complex topics over many minutes, synthesize findings from dozens of sources, and produce comprehensive reports.

Education example: Researching best practices for AI governance policies, analyzing peer institution approaches to enrollment challenges, or reviewing regulatory changes.

4. Voice Mode

Speak to AI naturally and hear spoken responses. This enables hands-free interaction and makes AI accessible in situations where typing is impractical.

Education example: Rehearsing a presentation, getting a verbal explanation of a concept while driving, or conducting a spoken brainstorm session.

5. Live Video and Screen Sharing

Share your screen or camera feed with AI and ask it about what it sees. AI can analyze visual content in real time and provide guidance.

Education example: Walking through a facilities issue on camera, getting feedback on slide design, or reviewing a physical document without scanning it.

6. Image Creation

Generate original images from text descriptions. AI can create illustrations, diagrams, infographics, and creative visuals on demand.

Education example: Creating visuals for newsletters, social media graphics, presentation illustrations, or event flyers.

7. Video Creation

Produce short video clips from text prompts or images. While still evolving, this capability is advancing quickly.

Education example: Creating brief promotional clips, animated explainers, or visual content for social media.

8. Autonomous Browser Control

AI can navigate the web independently, visiting websites, filling out forms, and gathering information without human intervention step by step.

Education example: Researching vendor pricing, gathering competitive enrollment data, or monitoring regulatory websites for updates.

9. Memory and Personalization

AI remembers your preferences, past conversations, and institutional context across sessions. Over time, it becomes more effective because it understands your specific needs.

Education example: AI that knows your district's terminology, your preferred report format, and your institutional priorities without being reminded each time.

10. Document and Image Analysis

Upload existing documents, images, charts, or screenshots and ask AI to analyze, summarize, or extract information from them.

Education example: Analyzing a scanned contract, summarizing a lengthy accreditation report, or extracting data from a photograph of a whiteboard.

11. Code Execution and Data Analysis

AI can write and run code to analyze data, create charts, and perform calculations. You do not need to know how to code; just describe what you want to understand.

Education example: Analyzing enrollment trends, creating visualizations from budget data, or processing survey results.

12. Projects

Create persistent workspaces where AI retains context, uploaded documents, and custom instructions across multiple conversations. Projects allow you to build a specialized AI assistant for a specific ongoing need.

Education example: A "Board Meeting Prep" project that holds your agenda templates, recent minutes, and institutional data so AI is always ready to help prepare materials.

13. Artifacts

AI can produce interactive content (calculators, dashboards, visualizations) that you can use, share, and modify. Artifacts go beyond static text to create functional tools.

Education example: Building a budget calculator, an interactive FAQ for students, or a data visualization dashboard.

14. Audio Generation

AI can generate spoken audio from text, including realistic voices, podcast-style content, and narration.

Education example: Creating audio versions of announcements, generating podcast-style summaries of board meetings, or producing narrated training materials.


Platform-Specific Strengths

While all three platforms share these 14 capabilities, each has areas where it excels:

  • Claude (Anthropic): Known for nuanced, detailed writing and strong reasoning. Claude Code is particularly powerful for technical and coding tasks.
  • ChatGPT (OpenAI): Broad general-purpose strength with deep integration into Microsoft products. Strong image generation with DALL-E and GPT image capabilities.
  • Gemini (Google): Deep integration with Google Workspace. NotebookLM is excellent for analyzing uploaded documents and generating audio summaries.

Choosing the Right Model Tier

Not all models within a platform are equal. Model selection deeply matters:

TierModelsCapability LevelBest For
A-GPT-5.2 Pro, Opus 4.5 Extended, Gemini 3 ProPhD Level (~90% GPQA, ~80% SWE)Complex analysis, strategy, nuanced writing
B-Sonnet 4.5, Gemini 3 ThinkingSr Professional (~83% GPQA, ~77% SWE)Most professional tasks, solid all-around
C-Haiku 4.5, GPT-5.2 Instant, Gemini 3 FastJunior Professional (~73% GPQA, ~73% SWE)Quick tasks, high volume, simple questions
FGPT 4o, Sonnet 3Intern (~53% GPQA, ~22% SWE)Basic tasks only; not recommended for professional use

GPQA measures graduate-level reasoning. SWE measures coding ability. Both benchmarks reflect the model's overall sophistication.

The key takeaway: using a top-tier model versus a bottom-tier model is the difference between working with a PhD-level professional and an intern. For important work, always select the best model available.


Getting Started

You do not need to master all 14 capabilities at once. We recommend this progression:

  1. Start with Text Chat (Capability 1): Get comfortable having conversations with AI
  2. Add Document Analysis (Capability 10): Upload real documents and ask questions about them
  3. Try Projects (Capability 12): Set up a persistent workspace for a recurring task
  4. Explore Deep Research (Capability 3): Use extended research for a real strategic question
  5. Experiment with the rest: Voice, video, image creation, and other capabilities as needs arise

The capabilities are available today, on platforms your institution may already have access to. The only barrier is learning to use them well.