Elicit Review 2026: Pricing, Accuracy Limits and Who It Suits

Elicit is an AI research assistant that reads academic papers rather than the open web. It searches a corpus of more than 138 million papers, pulls data from them into tables, and runs a dedicated systematic review workflow, which is why its industry pages target pharma, medtech, consumer goods and academia.

This review is for a researcher, analyst or graduate student deciding whether a paid Elicit plan beats the free tier or a cheaper rival. It covers what each plan costs on the two price lists Elicit runs, how its new usage pool works, what its accuracy claims rest on, and where a different tool fits better. Every price and feature comes from Elicit's pricing page and help center, read on October 21, 2026. We ran no extraction tests, so accuracy here is Elicit's own published evidence, labelled as such.

Dupple data: our AI research statistics page counted 18,327 arXiv papers ingested by Dupple's research radar in 2026 through August 19, from one preprint server alone. That volume is the problem Elicit sells against: no reviewer reads every abstract in a field that moves this fast.

Elicit is an AI research assistant built on academic literature. Its Research Agent answers a question from papers, Research Reports synthesize findings with sentence-level citations, and the Systematic Review workflow screens and extracts data from thousands of papers. It searches papers from sources including PubMed, OpenAlex, Semantic Scholar and Springer, and its Plus plan adds search across more than 500 thousand clinical trials. It also accepts uploaded PDFs and Zotero, EndNote, Mendeley and BibTeX libraries.

It is not a general chatbot. Elicit's help center says it has focused on empirical research, such as randomized controlled trials in social sciences and biomedicine, so questions outside that shape get weaker results.

One usage pool per month. Elicit replaced its old per-workflow counts and daily Research Agent caps with a single monthly pool shared across every feature on the plan, according to its usage limits article. A quick question draws little from it; a long agent session, a large review or figure extraction draws much more, and unused usage does not roll over.

Extra usage on Pro and Scale. Paid users who run out can switch on extra usage with a monthly spend limit they set. Elicit bills it in $100 increments as it is used, so a heavy month can cost more than the seat. When the pool is empty, Find Papers in the Research Agent still works on every plan, so a lapse slows a review rather than stopping it.

Two price lists. A dropdown on the pricing page switches between Industry and Academic plans, and a toggle switches between monthly and yearly billing, per Elicit's billing guide. The academic list is cheaper at every tier and is the only one with Plus, so anyone at a university should switch the dropdown before comparing Elicit with a rival.

Elicit's pricing page, read on October 28, 2026, shows the industry list by default with yearly billing. Pro is $49 per user a month on yearly billing, which Elicit says saves 35% against monthly billing, and Scale is $229 billed as $2,748, a 39% saving. The academic list prices Plus at $11 (billed $132 a year), Pro at $39 ($468) and Scale at $89 ($1,068). Enterprise is a custom contract. Check the current monthly price on Elicit's site before committing to a year.

The academic discount is largest on Scale. An academic Scale seat costs $1,680 a year less than the industry seat, while academic Pro saves $120, our arithmetic on the printed yearly charges. A three-person academic team on Scale pays $3,204 a year against $8,244 on the industry list.

Against rivals, Elicit is the expensive specialist. A year of industry Pro costs about four times the $144 yearly charge for Consensus Pro, and slightly more than the $540 yearly charge for Consensus Deep, the plan built for frequent literature reviews. The gap buys the systematic review workflow and extraction tables, so it is worth paying only if you use them.

Usage is the hidden variable. Every paid plan shares one monthly pool and extra usage bills in $100 steps, so budget for a heavy month, not only for the seat.

Elicit fits researchers who run structured literature reviews: pharma and medtech evidence teams, health economists, policy analysts and graduate students writing a systematic review. The extraction tables and screening workflow replace weeks of spreadsheet work, and the citations make every cell checkable.

Skip Elicit if you mainly want quick answers with sources, where Perplexity or the tools in best AI search engines cost less, or if you want to chat with your own documents, where NotebookLM, now called Gemini Notebook, and the options in NotebookLM alternatives fit better. Students on a budget should start with the free tier and the picks in best AI tools for students.