The top 10 fiction books of 2025, according to 91 lists

The Top Fiction of 2025, According to 91 Lists

7 min readJan 5, 2026

--

Happy New Year! As in the past six years, I’ve compiled a bunch of end-of-year “Best Books” lists from across the web to see which titles get mentioned the most. Personally, I’ve found this a great source of recommendations for my own reading, since bubbling to the top across the breadth of lists I scrape indicates fairly universal appeal: book critics, TikTok, and Obama agree: these are just good books.

This year, I sourced 91 lists from across the web. Here’s what showed up most often:

Read more about each at Goodreads:

1. The Loneliness of Sonia and Sunny by Kiran Desai (37 lists)
2. Heart the Lover by Lily King (34 lists)
3. Audition by Katie Kitamura (30 lists)
3. Flesh by David Szalay (30 lists)
5. Flashlight by Susan Choi (29 lists)
6. A Guardian and a Thief by Megha Majumdar (28 lists)
7. What We Can Know by Ian McEwan (27 lists)
8. The Emperor of Gladness by Ocean Vuong (22 lists)
8. The Wilderness by Angela Flournoy (22 lists)
10. Wild Dark Shore by Charlotte McConaghy (20 lists)
10. The Antidote by Karen Russell (20 lists)
12. We Do Not Part by Han Kang (19 lists)
12. Atmosphere: A Love Story by Taylor Jenkins Reid (19 lists)
14. The Buffalo Hunter Hunter by Stephen Graham Jones (17 lists)
14. Isola by Allegra Goodman (17 lists)
14. Buckeye by Patrick Ryan (17 lists)
16. Katabasis by R. F. Kuang (16 lists)
16. Dream Count by Chimamanda Ngozi Adichie (16 lists)
16. Bury Our Bones in the Midnight Soil by V.E. Schwab (16 lists)

Honorable mention to A Marriage At Sea by Sophie Elmhirst, a nonfiction book that broke into my list with 19 list mentions even though I didn’t scrape any non-fiction only lists (and plenty of fiction-only).

Skeptical? Don’t want your least-favorite news source taking up space? You can select just the lists you want here. That link will also let you browse books beyond the Top 19.

A note on accuracy, as I’ve done in the past as well: which lists are the best predictors of top fiction, or put another way, which lists are the most accurate? There are a few ways to look at this:

Precision: which lists have entries most likely to be in the Top 19? (# of Top 19 entries divided by list length)

Short lists have an advantage in precision: if you just named your favorite book and it happens to be on the Top 19, you have 100% precision!

NPR’s Fresh Air list from book critic Maureen Corrigan is a good example of this: she listed 10 books, 4 of which were fiction, and all 4 are in the Top 19, so they have 100% precision (4 of 4).Taking a fiction book rec from Fresh Air gives you a high chance of a “good book”, but you’ll run out of books quickly.

As usual, Obama is in the Zeitgeist.

Recall: which lists mentioned the most Top 19 books? (# of Top 19 entries divided by 19)

Long lists have a similar advantage in recall as short lists in precision: if you listed all books published in 2025, you would have 100% recall , because your list would mention every “Top” book — but your precision would be quite low. The top 5 highest-recall lists are all at least 30 books long, and most by definition can’t hit over 50% precision (maximum 19 “Top 19” books).

For example, if you read all 66 fiction books on Libby Life’s list, you’ll eventually get to almost all of the very top ones (17 of 19), but it’ll take you awhile to get through all 66 books on the list.

Accuracy: so, how to combine precision and recall? One way is to use something called the F-score. I’ll let wikipedia explain the math.

Similar to Precision, this still favors shorter lists in my dataset, but you’ll notice that the shortest ones (like NPR Fresh Air) disappear due to their lower recall.

Past years:

A note on tools: this year, I largely used Gemini to extract book titles and authors from articles. In the past, I’ve used webscraper.io, a great little Chrome Extension, to scrape well-formatted websites (for example, when the title is always in a particular font style, and the author in another). Unfortunately, not all lists are perfectly formatted — some are even in prose — and relying on formatting can become quite slow, with lots of double-checking and counting needed.

After some very thorough QA on my first few attempts, I found the Gemini 3 Flash model to be completely accurate at extracting the list from copy-pasted text, and almost perfect at assigning genre from {fiction, non-fiction, poetry} as well. This reduced my processing time significantly, once my accuracy concerns were assuaged. I still did a lot of spot-checking throughout, but even when unintuitive results came up (like Ali Hazelwood having THREE separate books on Goodreads Top 100 Most Popular), this process was much faster than in previous years. I was able to scrape lists that I’d skipped in the past that were written in prose and I previously couldn’t be bothered to parse.

Then, as in past years, I put everything in a Google spreadsheet and used Looker Studio to create my calculations and graphs — I’m not quite ready to outsource that to AI yet, but let’s check in next year.

Finally, the list of lists, for the curious.

The Top Fiction of 2025, According to 91 Lists

Written by Kimberly Johnson