Best AI for studying
For studying, Gemini Notebook, the product that was NotebookLM until recently, comes first today: it turns one set of sources into nine output formats and holds fifty sources per notebook on the free plan. Persian sits on its output-language list too, so an English source can come back as a Persian study guide or audio overview.
- four tools, three of them ranked
- weights on the page
- Iran column with its method stated
Last checked: This is a reference page. It is re-checked against the vendor sources and updated when a new version ships. What changed
The ranking today, from what each vendor published itself
A tool whose numbers we could not verify still appears in the table without a rank. Leaving it out would have made the table tidier and less true.
| Rank | Tool | Score | Ways to digest 35 | Your own material 20 | Workspace 20 | Output languages 10 | Access from Iran 15 |
|---|---|---|---|---|---|---|---|
| 1 | Gemini Notebook Current pick | 100.0 | 4/4 Nine output formats from one set of sources, two of them flashcards and quizzes, which is active recall rather than rereading | 50 sources Source | 3/4 Persistent notebooks, a hundred of them on the free plan, sharing with up to fifty people, and a mobile app, but no delivery or publishing workflow | 80 languages Source | blocked / no working route Google own help says Gemini Notebook works in the same 180+ regions where the Gemini app works, and Iran is not on the supported-countries list for the Gemini apps. We read that on Google page rather than measuring it; a direct test from inside Iran has not been run yet. |
| 2 | Gemini | 38.3 | 3/4 It builds study guides, flashcards and quizzes with charts, and pulls in YouTube videos and diagrams, but Google own help says audio overviews, video overviews and infographics stay in Gemini Notebook | 10 sources Source | 2/4 Chats and notebooks persist in the sidebar and sync with Gemini Notebook, but a shared notebook is not visible in Gemini at all and there is no team workspace | 70 languages Source | blocked / no working route Google own page says the Gemini web app runs in over 230 countries and territories, and Iran is not on that list. We read it there rather than measuring it. |
| 3 | Claude | 35.0 | 1/4 It produces text and documents, and its learning mode asks questions instead of answering, but it ships no audio overview, no video, no flashcards and no quiz format | not verified | 3/4 Unlimited projects on the Pro plan to keep chats and documents organised, team plans, and Canvas and Panopto connections in the university edition | not verified | blocked / no working route Iran is on neither of Anthropic two supported-countries lists. We read that on Anthropic own page rather than measuring it. |
An empty cell means we could not verify that number, not that the tool scored zero.
Not enough verified data to rank
These are real tools we track, but more than half of their criteria have no verified number yet, so ranking them would be a guess.
- ChatGPT
How this ranking is calculated
Every criterion below has a weight and a source. Change a weight and the whole table recomputes. There is no hand-placed position anywhere in this hub.
| Criterion | Weight | Evidence |
|---|---|---|
| Ways to digest | 35 | defined scale, with a written reason per assignmentOne source coming back in several shapes: summary, audio, video, mind map, flashcards, a quiz. Active recall is what moves learning, not rereading |
| Your own material | 20 | vendor stated specificationThe number the vendor published itself, and it does not count the same thing for every tool. The explanation is further down this page |
| Workspace | 20 | defined scale, with a written reason per assignmentWhether sources and outputs stay somewhere you can return to next month, or every session starts from nothing |
| Output languages | 10 | vendor stated specificationFor someone studying in Persian, the output-language list matters more than model quality |
| Access from Iran | 15 | our access column, with its method statedAll three ranked candidates are identical in this column today, so it separates nobody and only records the situation |
Why Gemini Notebook leads
Two numbers, both from Google own help pages. First, it turns one set of sources into nine outputs: notes, audio overviews, video overviews, mind maps, reports, data tables, flashcards or quizzes, slide decks and infographics. Second, the free plan holds fifty sources per notebook, each up to five hundred thousand words. The Gemini app, by comparison, takes ten files per prompt: enough to ask a question, not enough to work through a course.
A third thing has no column in the table and matters more than either for a Persian reader: Persian is on Gemini Notebook output-language list, and the same help page says the chosen language applies to study guides, audio overviews and chat responses. Give it an English source, get Persian back, with no trick involved.
What this table does not measure
Answer quality. All we have measured here is how many reviewable shapes a tool gives your material back in, and how much of that material fits. Claude sits third and may well reason and write better than both tools above it; it ranks low because it ships no audio overview and no flashcards, and Anthropic publishes no number at all for how much a project holds. An empty cell in our table means the vendor published nothing, not zero.
One note about the 100 itself. Normalisation is relative, and in each column the best candidate takes that column full weight. So 100 means first in all five columns of a four-candidate table, not perfect. Add a tool tomorrow that holds a hundred sources per notebook and that same score drops by itself.
Why ChatGPT carries no numbers
Because we have read nothing from OpenAI. Pages on openai.com return 403 to our server, both through our own fetcher and through plain curl, and the one domain that does answer is the API documentation, which says nothing about study mode. A number quoted on someone else blog is not a source. So the name is in the table, its cells are empty, and the day those pages open they get filled here.
What building this table taught us
Google renamed the product. The official page, the browser title and the support centre all say Gemini Notebook, while nearly every Persian search still says NotebookLM. If you are hunting for a tutorial and finding nothing, that is why.
The second thing, for anyone who wants to check us: the numbers live only on the plan and upgrade pages. Across all three companies the product marketing page carries not one printable figure. Google at least publishes a limits table; Anthropic publishes no capacity number at all, and that difference in publishing, rather than a difference in capability, explains part of the gap in this ranking.
When not to take our first pick
- If your work is writing and argument rather than digesting a pile of sources, ignore this table. Claude comes third here and is not a bad pick for that.
- If you only have one question and want the answer, building a notebook is overhead. The chat app is enough.
- If your source is a photo of handwritten notes, check the extracted text first, or all four output formats get built on top of a bad transcription.
Where it falls short
- Two of the five criteria are ordinal scales, which means our judgement with a written reason. We found no published numeric leaderboard for this job.
- The material column does not count the same thing for each tool: fifty sources in a persistent notebook and ten files in one prompt are not addable.
- ChatGPT is unmeasured, because OpenAI pages return 403 to our server.
- None of the three ranked candidates has an official route from Iran, so that column separates none of them today.
- Khanmigo, Perplexity and Quizlet have not been assessed yet. Their absence means we have not got to them, not that they failed.
Our take
If you have notes, books and papers and want to review those, take Gemini Notebook and set the output language on day one. If your work is writing rather than reviewing, take Claude. And if week one went into flashcards and podcasts without a single quiz, the tool did its job and you have not studied yet.
Questions people actually ask
What is the best AI for studying
Today, Gemini Notebook, because it turns one set of sources into nine output formats and holds fifty sources per notebook. That answer carries a date and the criteria behind it are on this page.
Is NotebookLM the same as Gemini Notebook
Yes. Google renamed the product to Gemini Notebook, and the official site and the help centre both use that name. The notebooklm.google address still works and lands in the same place.
Why does ChatGPT have no numbers in the table
Because no page on openai.com opens for us; every request comes back 403. We do not print a figure we have not read at its source, even where someone else has quoted it.
Can these be used from Iran
Not officially. Iran is not on Google supported-countries list for the Gemini apps, nor on Anthropic list. We neither sell accounts nor recommend circumvention.
If you want the same thing at the scale of a website, raw sources turned into publishable text, that work is part of our specialist content service