THE AI FIELD GUIDE

Find the right AI for the job.

Models change. Features move. Compare the evidence, see what each platform can actually do, and put it to work.

WHAT’S NEW · UPDATED NOV 10, 2026

What changed in AI tools and models

Compare models →

7 new and 15 updated in the last 30 days. “New” dates are the provider’s release dates; “Updated” dates are when the provider made the change or when we updated this guide.

  • New

    Claude Sheets, Whiteboards and Claude Data

    Sheets and Whiteboards (beta) arrive on paid Claude plans, and Dashboards becomes Claude Data.

  • New

    Claude Fable 5.2

    Anthropic’s top model is up to 30% faster and cheaper at $8 / $40, and now leads the Artificial Analysis index at 65.

  • New

    Muse Ember

    Meta’s low-cost model for agent sub-tasks and large batches, at $0.10 / $0.50 per million tokens.

  • New

    Claude Personal Finance

    Read-only answers about spending from connected US bank and card accounts, on Pro, Max and Ultra.

  • New

    GPT-6 Cyber and Claude Mythos 5.2 ↗

    Security-focused models for approved organizations only (OpenAI Daybreak Red; Anthropic’s verification programs), so they are not in the model catalog.

  • New

    GPT Image 2.6 Sunburst

    Images 2.6 brings faster generation, animated GIFs and SVG export to ChatGPT, Codex and the API.

  • New

    GPT Image 2.6 Flare

    The faster Images 2.6 variant, from about six seconds per image at low quality.

  • Updated

    Gemini 4 Carbon

    No longer a restricted preview: generally available on the paid Gemini API and rolling out to Google AI Ultra in the Gemini app, AI Mode and Antigravity.

Show 14 more changes
  • Updated

    Claude plans

    Max is $100 for 5x Pro usage; the $200 tier is now Ultra, with 10x weekly usage and no five-hour limit.

  • Updated

    Claude Code plans

    Max and Ultra replace Max 5x and Max 20x; both include Fable for up to half of weekly limits.

  • Updated

    Claude Artifacts

    Sheets, Whiteboards and Claude Data join documents and decks beside the conversation.

  • Updated

    Gemini

    Google AI Ultra subscribers now get Gemini 4 Carbon in the Gemini app.

  • Updated

    Google Antigravity

    Voice mode, a Knowledge Work layout for non-coding work, and Carbon with the /swarm multi-agent preview on Ultra.

  • Updated

    Kimi K3.1 Pro

    Its promised full open weights are now on Hugging Face.

  • Updated

    Kimi

    Kimi K3.1 Pro’s full weights are now published.

  • Updated

    Reasoning & knowledge leaders

    Artificial Analysis v5.0.1 regraded one test; Claude Fable 5.2 now leads at 65, ahead of Gemini 4 Carbon at 60.

  • Updated

    Web development leaders

    Claude Opus 5.5 still leads Arena’s voted WebDev board. Fable 5.2 has only an automated, unranked result so far.

  • Updated

    Muse Code

    Muse Ember is included on every plan.

  • Updated

    Writing & conversation leaders

    Gemini 4 Carbon and Argon lead Arena’s text board; Muse Wonder moves to third.

  • Updated

    Image generation leaders

    GPT Image 2.6 enters in second place on early votes, behind GPT Image 2.5 Sunburst.

  • Updated

    Video generation leaders

    Gemini Omni 1.2 Flash, released November 5 with 15-second clips and 30% lower prices, leads on preliminary votes.

  • Updated

    ChatGPT

    Images 2.6, with animated GIFs, on every plan.

WHO’S AHEAD?

Leaders by task

Compare benchmark results →

Leading models by task, with scores and sources.

WHAT IT COSTS

ChatGPT, Claude, Gemini and the rest: free tiers and prices

Most people should start on a free tier and pay only when a limit gets in the way. Coding tools are compared further down.

AI assistant plans. US list prices before tax, checked October 22, 2026 on each vendor’s pricing page.
ToolFree tierPaid plans
ChatGPT Updated GPT-5.6 Luna, with Think for harder questions, and quick Codex tasks. No GPT-6.
  • Go $8/mo
  • Plus $20/mo
  • Pro $100, $200 or $500/mo
  • Business $25/user/mo ($20 yearly)
Claude Updated Sonnet and Haiku, web search, file creation and code, memory, connectors, artifacts, Claude Design, Docs and Slides, up to 5 Projects and voice, with the lowest usage limits. No Claude Code or Opus.
  • Pro $20/mo ($17 yearly)
  • Max $100/mo
  • Ultra $200/mo
  • Team $25/seat/mo ($20 yearly)
Gemini Updated Flash-Lite model with Canvas, Deep Research, Gems and Skills, Gemini Live, image generation and Gemini Notebook, at base usage limits.
  • Google AI Plus $4.99/mo
  • Google AI Pro $19.99/mo
  • Google AI Ultra $99.99 or $199.99/mo
Microsoft Copilot Copilot chat on the web, Windows and mobile at base usage limits.
  • Microsoft 365 Personal $9.99/mo
  • Microsoft 365 Family $12.99/mo
  • Microsoft 365 Premium $19.99/mo
  • Microsoft 365 Pro $99.99/mo
  • Microsoft 365 Copilot Business $18/user/mo yearly
Grok Daily Grok access, a starter allowance for Cursor’s coding agents and standard X.
  • Plus $8/mo
  • Premium $30/mo
  • Super $100/mo
  • Ultra $200/mo
Perplexity Cited web answers; model choice, Deep Research and higher limits need Pro.
  • Pro $17/mo billed yearly
  • Max $167/mo billed yearly
01 / MODELS

What’s under the hood?

Compare capabilities, access, API rates and reference scores side by side.

Choose up to three →
02 / PLATFORMS

More than a chat box.

Dots, Sites, Artifacts, Skills, Grok Build and the features a model score won’t tell you about.

See what you can do →
03 / GUIDES

Start with a real task.

Practical workflows for research, coding, media and everyday work.

Try a guide →

Tools & platform features

34 results

OpenAI

ChatGPT

Updated

OpenAI’s general-purpose assistant for writing, analysis, coding help, image generation (Images 2.6, with animated GIFs, since November 4) and voice, on web, desktop and mobile. Plus and above use GPT-6; Free and the $8 Go plan use GPT-5.6 Luna.

Useful for: A broad all-rounder with the largest ecosystem: Codex for coding, Sites for publishing, Dots for ongoing work and plugins for other apps.

Explore ChatGPT →
Anthropic

Claude

Updated

Anthropic’s assistant for writing, analysis, research and coding. Cowork tasks, Docs, Slides and artifacts live in one Claude conversation, and since November 10 so do Sheets, Whiteboards and Claude Data.

Useful for: Long documents, careful writing and coding, and turning a conversation into a document, deck or dashboard.

Explore Claude →
Google

Gemini

Updated

Google’s assistant on the web, Android, Windows and Mac, and inside Gmail and Docs. Gemini 4 Argon reached Pro and Ultra subscribers on October 13, and the more capable Gemini 4 Carbon reached Ultra subscribers on November 10.

Useful for: People who live in Google’s apps and Android, and anyone who wants a capable free tier with Canvas and Deep Research included.

Explore Gemini →

Microsoft’s assistant in Windows, Edge and Microsoft 365 apps such as Word, Excel, Outlook and Teams.

Useful for: Households and businesses already paying for Microsoft 365, where Copilot works inside the documents and email they use every day.

Explore Microsoft Copilot →
Perplexity AI

Perplexity

An answer engine that searches the web and responds with cited sources, with a choice of models (GPT-6.1 Sol, Claude Sonnet 5.5, Gemini 3.8 Flash, Grok 4.7, Kimi K3) and deeper research modes.

Useful for: Quick researched answers with sources you can check, and trying several companies’ models under one subscription.

Explore Perplexity →
SpaceXSI (formerly xAI)

Grok

SpaceXSI’s assistant for chat, voice, file analysis and image and video generation. Since October 12 it is sold with Cursor and X Premium as The X Subscription.

Useful for: People who also use X or Cursor, since one plan now covers all three, and anyone who wants image and video generation in the same app as chat.

Explore Grok →

Meta’s personal AI agent, launched September 8, that does tasks such as sending email, booking travel and filling in forms on its own secure cloud computer. You can message it in the Muse app or in WhatsApp.

Useful for: US users who want an agent that completes errands for them, inside apps they already use such as WhatsApp.

Explore Meta Muse →
DeepSeek

DeepSeek

A chat app and low-cost API whose V4.1 Flash model handles text, vision, reasoning and tool calls with a 1M-token context.

Useful for: Developers who want a cheap API for tasks with checkable answers, especially batch work that can run in off-peak hours at half price.

Explore DeepSeek →
Anthropic

Claude Code

Updated

Anthropic’s coding agent that reads, edits and runs code in your project, from the terminal, VS Code and JetBrains, the desktop and mobile apps, the web, Slack and GitHub Actions.

Useful for: Delegating multi-step coding tasks across a codebase, from wherever you work.

Explore Claude Code →
Cursor (SpaceX)

Cursor

An AI-first code editor built on VS Code, with codebase-aware chat, multi-file edits and local and cloud agents. Part of SpaceX since August, and sold with Grok and X Premium as The X Subscription.

Useful for: Developers who want AI woven through the whole editor, especially those who will use the included Grok models.

Explore Cursor →
GitHub (Microsoft)

GitHub Copilot

Code completion, chat and agents inside VS Code, JetBrains IDEs, GitHub and the command line, with models from Anthropic, OpenAI, Google, SpaceXSI and Moonshot.

Useful for: Developers who want AI help inside the editor they already use, with a choice of models on one bill.

Explore GitHub Copilot →
OpenAI

Codex

OpenAI’s coding agent in the ChatGPT apps, the command line, IDE extensions, the web and iOS, with cloud environments and a security-review preview.

Useful for: ChatGPT subscribers who want a coding agent included in a plan they already pay for.

Explore Codex →
Google

Google Antigravity

Updated

Google’s agentic development platform: a desktop app for running agents in parallel, an IDE, a CLI and IDE extensions, also built into the Gemini Mac app. Since November it has voice mode and a Knowledge Work layout for research, documents and data work.

Useful for: Developers who want a free coding agent, or Google AI subscribers who want Gemini 4 Argon for coding.

Explore Google Antigravity →

Google’s research notebook, renamed from NotebookLM in July 2026. It answers questions grounded in the documents, links and videos you add, and makes Audio and Video Overviews, slides and flashcards.

Useful for: Studying or summarizing your own source material, with answers that point back to the passage they came from.

Explore Gemini Notebook (NotebookLM) →
Midjourney

Midjourney

An image generator known for its distinctive, high-quality artistic output, with video generation.

Useful for: Stylized illustration and concept art.

Explore Midjourney →
Moonshot AI

Kimi

Updated

An assistant and API from Moonshot AI. Kimi K3.1 Pro is the latest K3.1 release for vision, coding and long-context work, and its full open weights are now on Hugging Face.

Useful for: Testing long-document and coding work against other model families. K3.1 Pro and its Highspeed variant have separate API offerings.

Explore Kimi →
Ollama

Ollama

Run open models on your own computer for free, or use Ollama’s cloud for larger models such as DeepSeek V4, Kimi K3 and gpt-oss.

Useful for: Keeping data on your own hardware, working offline, or connecting open models to your own apps without per-token bills.

Explore Ollama →
Meta

Muse Code

Updated

Meta’s multi-agent coding tool for the terminal and CI, built for its Muse models, with approvals and an OS sandbox on from the first run. It is the only way to use Muse Wonder Full.

Useful for: Developers who want a low-cost coding subscription, or access to Muse Wonder Full, which is not sold through the API.

Explore Muse Code →
Replit

Replit

A browser-based development environment with an AI agent that can build, host and deploy apps from a description.

Useful for: Prototyping and deploying small apps without local setup, including for people who do not write code.

Explore Replit →

Adobe’s generative image, video and audio tools, in the Firefly app and in Photoshop, Illustrator and Premiere, with partner models from Google, OpenAI and others.

Useful for: Creative Cloud users who want generative tools inside Adobe apps, and trying several image models with one set of credits.

Explore Adobe Firefly →
Runway

Runway

AI video generation and editing for creators and production teams, with Runway’s Gen-4.5 and Aleph models alongside partner models such as Seedance and Kling.

Useful for: Generating and editing short video clips, and comparing several video models in one app.

Explore Runway →
Notion

Notion AI

Notion Agent, AI meeting notes and search across your pages and connected apps, inside Notion workspaces.

Useful for: Teams that already keep their docs and wikis in Notion.

Explore Notion AI →
Superhuman (formerly Grammarly Inc.)

Grammarly

Writing assistance for grammar, tone and rewriting across browsers, email and documents, now part of the Superhuman suite.

Useful for: Polishing everyday writing wherever you type.

Explore Grammarly →
Otter.ai

Otter.ai

Meeting transcription, summaries and action items for Zoom, Google Meet and Microsoft Teams.

Useful for: Automatic notes from recurring meetings.

Explore Otter.ai →
ElevenLabs

ElevenLabs

Realistic text-to-speech, voice cloning, dubbing, music and speech-to-text, through an app and API.

Useful for: Narration, dubbing and voice features in apps.

Explore ElevenLabs →

Build, host and share a website or small web app by describing it in ChatGPT or Codex, including sites that use each visitor’s own connected apps.

Useful for: Publishing an event page, small business site or interactive web tool from a written brief.

Explore ChatGPT Sites →

Always-on agents powered by GPT-6 Astra, with their own cloud computer and browser, that keep ongoing work moving and ask when a decision needs you.

Useful for: Recurring research and follow-up work with a clear scope and a useful reporting schedule.

Explore ChatGPT Dots →
Anthropic

Claude Artifacts

Updated

Documents, decks, sheets, whiteboards, dashboards, code and small interactive tools that open beside a Claude conversation, which you can edit, revisit and share with your team.

Useful for: Turning an explanation into an editable document, visualization or small interactive tool.

Explore Claude Artifacts →

A Gemini workspace for drafting documents and building apps, quizzes, dashboards and infographics, with export to Docs, Slides or Colab.

Useful for: Revising a document or interactive prototype while keeping the draft in view.

Explore Gemini Canvas →

An agent that plans and runs a research task across up to hundreds of websites, and optionally your Gmail, Drive and Chat, then writes a report.

Useful for: Research questions that need several sources, a date range and an explanation of conflicting evidence.

Explore Gemini Deep Research →
SpaceXSI (formerly xAI)

Grok Build

SpaceXSI’s coding agent, with an open-source terminal interface that can run against local models, plus web and mobile versions and memory that carries across sessions.

Useful for: Delegating a bounded repository change through a terminal coding workflow.

Explore Grok Build →
SpaceXSI (formerly xAI)

Grok Bot

Always-on agents with their own cloud computer that use apps, browse and run code, with routines, triggers and a review agent that checks proposed actions.

Useful for: Longer-running research or browser work with a defined deliverable and clear action limits.

Explore Grok Bot →
SpaceXSI (formerly xAI)

Grok Imagine

Image and video generation in Grok. Imagine Image 2.0 handles precise edits; Imagine Video 1.5 takes text, image and voice references and outputs up to 1080p.

Useful for: Trying image concepts and short video ideas from a visual brief.

Explore Grok Imagine →

Coding tools: Claude Code, Codex, Copilot, Cursor, Antigravity and Muse Code

Agent work uses allowances quickly. Compare what a finished task costs you, not the headline price.

AI coding tool plans. US list prices before tax, checked October 22, 2026 on each vendor’s pricing page.
ToolFree tierPaid plans
Claude Code Updated Not included on the Free plan.
  • Pro $20/mo ($17 yearly)
  • Max $100/mo
  • Ultra $200/mo
  • Console API rates
Codex Quick coding tasks; Go ($8) adds lightweight tasks.
  • Plus $20/mo
  • Pro From $100/mo
  • Business $25/user/mo ($20 yearly)
GitHub Copilot 2,000 code completions a month and limited chat and agent use. No cloud coding agent.
  • Pro $10/mo
  • Pro+ $39/mo
  • Max $100/mo
Cursor Limited agent and chat requests.
  • Plus $8/mo
  • Premium $30/mo
  • Super $100/mo
  • Ultra $200/mo
  • Teams From $30/user/mo
Google Antigravity Updated No charge to start: unlimited tab completions and commands, basic weekly agent limits and some agent models.
  • Google AI Pro $19.99/mo
  • Google AI Ultra $99.99 or $199.99/mo
Muse Code Updated No free subscription is listed; you can pay per token through the Meta Model API instead.
  • Everyday Usage $5/mo
  • High Usage $15/mo
  • Power Usage $50/mo
  • Maximum Usage $200/mo

Make it useful

All guides →

Latest AI news

Image via The Verge
The Verge

Trump’s latest land grab is an AI hub in the Philippines ↗

President Donald Trump was greeted by no less than five beauty queens when he landed in the Philippines on Tuesday, according to local news outlet Rappler. As he stepped onto the tarmac, a child handed him small US and Philippine flags. A band played "YMCA," a staple at Trump rallies, as more girls in traditional…

Excerpt · The VergeRead at The Verge ↗
All AI news →