How to Read an IMDb Rating
What a weighted average actually measures, why a 7.4 is not the same as a 7.4, and how to sanity-check any score before you commit an evening to it.
Read article →Independent film & TV review site. We publish criticism and information only. Nothing on this site is downloadable and we host no software or install files. Read our disclaimer
Independent film & TV review site. We publish reviews and information only. Nothing on this site is downloadable and we host no software or install files. Read our disclaimer
The best documentaries of 2026 so far, ranked and scored on our own ten-point rubric, with honest notes on the access each film got and what it did with that access.
Documentary is the only form where the making of the thing is part of the review. A fiction director invents a world and you judge the invention. A non-fiction director walks into a world that already exists, talks a door open, and then has to decide what to do with whatever is standing behind it. That negotiation is usually invisible in the finished film, and it is the first thing we weigh when we score one. This is our ranking of the ten best documentaries of 2026 so far, from ten to one, on the same ten-point editorial scale we use everywhere else on this site.
Our published rubric grades six categories, and three of them carry more weight for a documentary than they would for a drama. Evidence is the first. Does the film show you the thing it claims, or does it show you somebody describing the thing? Access is the second, and access is not the same as intimacy: a camera allowed into a room for six weeks is worth more than a camera allowed in for one arranged afternoon, and you can usually feel which one you are watching. The third is what the film does with the access once it has it. Sitting in a room is not a thesis.
The categories that matter less here are the ones that dominate fiction scoring: performance, narrative shape and production polish. A documentary can be beautifully shaped and still be hollow, and it can be shapeless and still change what you understand. We marked down gorgeous films this year for telling us nothing we could not have read in a press release, and we marked up rough ones for getting somewhere nobody had been allowed to stand before.
One rule before we start. Everything on this list had a licensed release, or a festival run with a confirmed distributor attached, between January and the end of September 2026. We do not rank a film you cannot legally watch. Where a title is streaming now we say so, and where it is still between festival and release we say that instead.
Ask of any documentary: if the subject had refused the camera, would the film still exist? If the answer is no, you are watching a portrait. If the answer is yes, you are watching an argument. Most of this year’s best work is the second kind, and it is not a coincidence that the second kind is also the kind that is hardest to get made.
10. Salt and Standing Water, 7.3 / 10. Marisol Vega’s film about a Louisiana parish slowly losing ground to subsidence is the most visually striking entry on this list and, frustratingly, the least curious. Vega gets extraordinary access to three families watching their land sink, and the images of flooded cemeteries are unforgettable. What she does not do is press. Nobody in the film is asked a hard question about the buyout programme, the insurance maths, or why the parish kept issuing building permits for a decade after the water table turned. A 7.3 is not a bad score. It is the score for a film that photographed a crisis beautifully without interrogating one.
9. The Quiet Ledger, 7.6 / 10. Halina Borys spent two years inside a municipal water authority’s forensic audit, and the result plays like a thriller because the stakes turn out to be a city’s supply. Its strength is paperwork: redacted invoices, meeting minutes, a spreadsheet that becomes genuinely menacing once you understand what the columns mean. It loses ground for a middle hour that assumes we are as fascinated by procurement law as the auditors are. We nearly were. Not quite.
8. Ninety Seconds of Static, 7.8 / 10. Tomas Rybakov reconstructs a besieged city’s amateur radio network through the operators who kept it on the air, and it does something rare: it makes a technical subject emotional without cheating. The archive audio is the evidence, the testimony is context, and the film resists putting a score over the top of either. It ranks eighth because the final twenty minutes reach for a redemptive ending the material never earned, and the film is better than the ending it chose.
7. The Long Field, 8.0 / 10. Aoife Dunmore filmed one dairy family across four seasons on land that has been theirs for a century and is no longer profitable. There is no narration and no expert interviews, just work: milking, repairing, arguing about a loan in a kitchen. It is the least dramatic film here and one of the most affecting, because Dunmore trusts you to understand what a falling milk price means when you can see the herd standing behind the person saying it. The 8.0 reflects a last act that gets a little sentimental about tradition; the film is strongest when it is merely observant and lets you do the feeling.
6. Paper Boats, 8.2 / 10. Kenji Aoyama built a documentary around an archive of letters written by flood survivors and never delivered, and the decision to have actors read them aloud is the one that divides audiences. We think it works, mostly. The letters are the evidence; the readings give them a shape a page cannot. Aoyama is careful to mark which passages are verbatim and which are translated reconstruction, which is more honesty than most films of this type bother with. It sits at six because the reconstruction occasionally smooths over exactly the ambiguity that makes the originals so powerful.
5. The Interpreter’s Room, 8.4 / 10. Yara Haddad observes the interpreters who work an immigration tribunal, and it earns its place by refusing to make the obvious film. There are no triumphant rulings here and no villains; there is the exhausting craft of carrying somebody else’s worst day across a language barrier without changing its meaning. The access is remarkable, the ethics of that access are visibly on the filmmakers’ minds, and the film is scrupulous about not turning applicants into case studies. Half a point off for a recurring visual metaphor the material never needed.
4. Kingdom of Small Machines, 8.6 / 10. Petra Lindqvist spent a year in a repair economy where nothing is thrown away and everything gets diagnosed, and made the most quietly political film of the year. Every scene is a small negotiation between a technician who can fix a thing and a manufacturer that would rather sell a new one. Lindqvist’s structural insight is the reason it scores this high: she does not interview a single executive, and the film is stronger for it, because the absence does the arguing. It misses the top three only because the last third restates a point it had already proved twice.
3. Understory, 8.8 / 10. Olek Marchetti’s film about an old-growth forest is the one documentary this year that genuinely changed how we look at something familiar. Shot across three years with canopy rigs his crew built themselves, it treats a forest as a slow, competitive, networked organism rather than as scenery. The science arrives through the people doing it, in the field, mid-climb, and the film never condescends to either the researchers or the loggers who disagree with them. Where it gives up ground: a closing sequence that reaches for awe when it had already earned wonder.
2. The Last Ferry to Ardmore, 9.0 / 10. Ines Bardi’s portrait of an island in its final decade of habitation is the most humane film on this list and the hardest to shake. Bardi moved to the island, stayed two winters, and filmed the people who stayed with it: the last teacher, the ferryman, the woman who keeps the register of the dead. There is no narration at all, and yet by the end you understand the economics, the weather and the emigration without anyone explaining a thing. It is second rather than first on one criterion only. Bardi is so careful with her subjects that she sometimes declines to press them when the film, and not the person, is the thing that needs pushing.
1. What the River Kept, 9.3 / 10. Solomon Adeyemi’s film follows a forensic team working to identify remains recovered from a riverbank decades after a massacre, and it is the best documentary of 2026 because it understands that the investigation is not the subject. The subject is what certainty costs. Families are asked to give a DNA sample on the promise of an answer that may be a name, a grave, or nothing at all, and Adeyemi films those conversations at length, camera steady, no music underneath.
What separates it from everything above is how rigorously it handles its own access. When a family asks the team to stop filming, Adeyemi stops filming, holds the empty frame, and leaves the gap in the cut. When a match fails, the film does not cut away to soften it. There is a working method on screen here rather than a result, and the film is honest that the method itself sometimes causes harm to the people it is meant to serve.
One criticism is worth making. At two hours and twenty minutes it is longer than it needs to be, and the final stretch revisits testimony the audience has already absorbed. Even so, no other documentary this year carried this much responsibility and dropped it this rarely.
What the River Kept takes the top spot by holding two difficult things at once: it gives you the archive, the method and the results, and it refuses to pretend the process is clean. If you watch one documentary from 2026, watch this one, then watch Understory for an entirely different idea of what a camera should be allowed to see.
2026 has been an unusually good year for documentaries that earn their access instead of assuming it, and the top four here would have placed highly in any recent year. What holds the group back is discipline: five of these ten are fifteen minutes longer than they need to be, and the endings are where ambition most often outruns judgment. On the evidence of the first three quarters, the final one should be worth waiting for.
More original guides from the same testing desk.
What a weighted average actually measures, why a 7.4 is not the same as a 7.4, and how to sanity-check any score before you commit an evening to it.
Read article →Six thrillers that respect a two-night binge, ranked by how well they hold tension across a full season without padding the middle.
Read article →What each tier actually gets you, where the price gaps sit, and how to work out which one fits your household before you upgrade by accident.
Read article →