Apps that make music, pictures and video, edit photos, research, write, take meeting notes, build websites, translate and teach. What each costs, where it runs, what changed last, and who owns what you make.
133 tools
Songs, pictures, video clips, voices and 3D characters from a written description.
Songs and instrumentals from a written description.
Pictures from a written description.
Short video clips from words or a still image.
Text read aloud in natural voices, and copies of a voice.
Change photos and videos by saying what you want, and design graphics, slides and social posts.
Change a photo by describing the change.
Cut, fix and change video with AI help.
Search the web with sources, read long reports for you, and help with writing.
Answers from the web, with the sources linked.
Help drafting, rewriting and checking text.
General chat assistants for everyday questions.
Meeting notes, slides, spreadsheets, translation and transcripts.
Notes and summaries of calls and meetings.
Presentations from a prompt or a document.
Formulas, tables and analysis in spreadsheets.
Text, speech and conversation in other languages.
Transcripts, captions and dubbed video.
Make an app or website by describing it, and browsers that do tasks on the web for you.
Apps and websites from a description, no coding needed.
Browsers and agents that click through websites for you.
Tutors and study help, health, money and legal tools.
Tutors, homework help and study modes.
On Arena (formerly LMArena), people compare two results for the same request without knowing which model made which, and vote. Many of the apps above run these models inside them.
Arena, Oct 7, 2026 · 6,604,125 votes; 82 models
Arena, Oct 6, 2026 · 30,985,932 votes; 59 models
Arena, Sep 21, 2026 · 718,577 votes; 48 models
Arena, Sep 21, 2026 · 2,081,953 votes; 48 models
Arena, Aug 26, 2026 · 27,930 votes; 10 models
Scores are Arena's ratings from votes; a gap of a few points is within its margin of error. “Preliminary” means few votes so far. Model names are Arena's.
Independent tests of AI assistants answering questions about the news. Both are from 2025; we haven't found newer ones of the same kind.
“Collectively, they provided incorrect answers to more than 60 percent of queries.”
“with Perplexity answering 37 percent of the queries incorrectly”
“Grok 3 had a much higher error rate, answering 94 percent of the queries incorrectly.”
Columbia Journalism Review, Mar 6, 2025
“Almost half of all AI answers had at least one significant issue.”
“A third of responses showed serious sourcing problems.”
“22 public service media organizations in 18 countries – working in 14 languages – assessed how ChatGPT, Copilot, Gemini, and Perplexity answer questions about news and current affairs.”
EBU, Oct 21, 2025
Which underlying models make things up least when summarizing a document: AI models compared. More warnings: Watch out.
Read Oct 20, 2026. Some makers' pages only show prices once a page runs in a browser, and some kinds of tool need another round of research.
Compare AI modelsComing next: budgets and taxes; contracts and legal research.
Prices are each maker's US list price a month, billed monthly unless its note says otherwise. Where a maker is based comes from general knowledge for most tools, not from the maker's pages.
