Home / Blog / Explainers

Attack on Titan Waited Ten Years for Its Ending — and That Is Why It Has Zero Filler

Attack on Titan runs 89 episodes with 0 filler — 83 canon, 6 mixed, nothing invented. How a finished manga, seasonal breaks and a decade of patience produced the cleanest long adaptation of the streaming era, and what the waiting cost.

Every long-running anime on this site makes its viewers do the same quiet piece of arithmetic: how much of this show is the story, and how much is the wait for the story? For most series the answer is a fraction, and sometimes a small one. Naruto Shippuden asks you to sit through 203 filler episodes. Bleach asks for 163. Sailor Moon asks for 98 out of 200. And then there is Attack on Titan, which asks for nothing. Zero filler episodes in a run of 89. Not one anime-only arc, not one beach episode, not one clip show. Every episode either adapts the manga or carries a piece of it. That is the whole story of this article, and it is worth telling in full — because zero filler in a series this long is never an accident. It is the visible result of a production plan, and the plan had a price.

This article looks at the numbers, at the two decisions that made them possible, at the six mixed episodes that are the only judgement call in the whole run, and at what the audience paid for the cleanliness. Because the zero-filler story always has a second half, and Attack on Titan's second half — ten years from first episode to last — is as revealing as its first.

The numbers: 83 canon, 6 mixed, 0 filler

Attack on Titan ran from 2013 to 2023 for 89 episodes. Our classification of the full run is: 83 manga-canon episodes, 0 anime canon, 6 mixed episodes and 0 filler episodes. The filler share is 0.0% — one of only three series on this site with no filler at all, alongside Demon Slayer (0 in 63) and Yu Yu Hakusho (0 in 112). The story route — everything except the filler — is all 89 episodes, about 34 hours at 23 minutes an episode, against 34 hours for the complete run. Skipping the filler saves nothing, because there is nothing to skip. That is the entire dividend, and it is also the point: the list for this series is a list with nothing on it.

Put that beside the rest of the catalogue and the gap is startling. Across the 26 series and 6,884 episodes on this site, 1,610 episodes are filler — 23.4%, nearly one in four. The median series is 11.9% filler. Attack on Titan sits at the absolute floor of that distribution, and it does so while being a genuinely long series with a decade-long broadcast history, not a short one that never had time to develop a filler problem. At 89 episodes it is longer than Fullmetal Alchemist: Brotherhood (64) and Demon Slayer (63), and it sustains its zero rate across three studios' worth of scheduling, a change of production studio midway, and a final season that was itself split into parts. Length, on its own, does not create filler. The conditions around the length do.

The other numbers in the classification are just as telling. Zero anime-canon episodes: unlike Boruto, which built its show out of 179 supervised original episodes, or Black Clover, which used 21, Attack on Titan invented nothing that the story keeps. Six mixed episodes, against Naruto's 56 or Shippuden's 64: the production almost never padded a canon episode with original material, either. The classification is about as pure as a long adaptation gets — 83 episodes that adapt the manga, six that mostly do, and nothing that does not. If you want to know what a filler list looks like when there is nothing to list, this is it.

Why zero filler is so rare

To understand why zero is remarkable, it helps to remember what filler is for. As we explain in our article on why long-running anime has so much filler, filler is what happens when a weekly anime catches up with its manga and has to keep broadcasting anyway. Three conditions have to hold for that trap to close: a manga that is still being written, an anime adapting it faster than it is written, and a broadcast slot that cannot pause. The weekly era built its business on all three at once — a permanent slot, a manga a few years old, and an adaptation pace of roughly an episode a week — and the result is written across this site in hundreds of filler episodes per series.

Attack on Titan avoided all three conditions, and it avoided them deliberately. The production was seasonal from the start: it aired a season, stopped, and came back when there was more manga to adapt and more production time to do it justice. A season that ends can wait. A weekly slot that never ends cannot. That single structural choice — seasons with gaps instead of a permanent weekly appointment — is the difference between Attack on Titan's zero and Bleach's 163, and it is the same choice that gives Demon Slayer its zero in 63 episodes and My Hero Academia its 4 in 170. The seasonal model is the filler problem solved at the level of the schedule, before a single episode is written.

It is worth saying plainly that this is a change in the industry, not a change in the virtue of individual studios. The studios that made the weekly filler arcs of the 2000s were not lazier than the studios making seasonal shows now; they were working inside a system that demanded an episode every week, forever, from a manga that produced a chapter a week. The seasonal system demands something else — patience from the audience, and a willingness from the broadcaster to let a slot go dark for a year or more — and in exchange it returns an adaptation that never has to invent material. Attack on Titan is the clearest demonstration of that exchange, because it made it at the largest scale: a decade, 89 episodes, and not one invented episode.

The manga finished first

The second condition that made zero filler possible was the manga itself. Attack on Titan's manga, by Hajime Isayama, ran from 2009 and finished in 2021, two years before the anime's final episode aired in 2023. The anime's ending was not a guess, and it was not an invention: it was an adaptation of an ending that already existed, written by the manga's author, and the production knew it was coming for years before it arrived. That matters more than it might sound, because the hardest part of adapting an unfinished manga is not the middle — it is the ending. A weekly show that catches up with its manga mid-story must invent; a seasonal show adapting a finished manga must only wait.

The waiting is the point. The anime began in 2013, when the manga was four years in and far from finished, and it could easily have become a weekly show: the manga was popular enough, the first season was a phenomenon, and a permanent slot was there for the taking. Instead the production took breaks — a four-year gap between the first and second seasons, shorter gaps after that — and let the manga build its lead. Each season adapted a completed stretch of story, and the final season waited until the manga's ending was in hand. By the time the last episodes were made, there was no unanswered question about where the story went, because the manga had answered it. A production that knows its ending can pace itself to arrive there. A production that does not know its ending must fill whatever time appears.

Compare that with the weekly era. The original Naruto began adapting a manga only a few years old and caught up with it on screen; Bleach started in 2004 with a manga from 2001; Sailor Moon adapted a manga of just 60 chapters into 200 episodes. In each case the anime was chasing a story that was still being written, and the chase is what produced the filler. Attack on Titan reversed the chase: the story finished first, and the anime followed at its own pace. It is the same lesson Hunter × Hunter (2011) teaches from a different direction — that series had a decade-long head start on its manga and produced just 2 filler episodes in 148 — and the two cases together make the mechanism hard to miss. Give a production a finished or nearly finished manga, and filler disappears. Take it away, and filler is what fills the gap.

Seasons, not a weekly slot

The seasonal structure deserves a closer look, because it is doing more work here than simply creating gaps. A weekly series and a seasonal series do not just air differently; they are planned differently. A weekly series plans week by week, because it does not know how long it will run or when the manga will end, and its production pipeline is a treadmill: an episode every seven days, indefinitely. A seasonal series plans season by season, with a defined episode count, a defined arc to cover, and a defined end date — and then it stops. The planning horizon is the whole season, and the production can allocate its best animation, its best directors and its best episodes to the moments that matter, instead of spreading them thinly across an endless run.

You can see the result in the shape of Attack on Titan's broadcast history. The first season aired in 2013. The second did not arrive until 2017 — four years later, an unthinkable gap for a weekly show, and a routine one for a seasonal production that was waiting for manga and production capacity. The third season followed in 2018 and 2019, and the final season, which began in 2020, was itself split into parts across three years. To a weekly-era viewer that looks like chaos; to a seasonal-era viewer it is the normal rhythm of a show that airs when it is ready. Each return was an event, each season was complete in itself, and none of them needed a filler arc to bridge a gap, because the gaps were the plan.

There is a second, quieter benefit to the seasonal model, and it shows in the mixed-episode count. Weekly productions do not only invent whole filler episodes; they also pad canon episodes with original scenes to slow the adaptation down — the opening recap that eats two minutes, the reaction shots that stretch a fight, the side conversation that was never in the manga. That padding is what our mixed category captures, and the weekly series are full of it: Naruto has 56 mixed episodes, Shippuden 64, Bleach 41. Attack on Titan has 6. A seasonal production, adapting a defined stretch of manga at a defined pace, has no reason to stretch a scene, because it is not trying to slow down. It is trying to arrive on time.

The studio handover that changed nothing

One of the most telling facts about Attack on Titan's zero-filler record is that it survived a change of studio. The first three seasons were produced by Wit Studio; the final season was produced by MAPPA, after Wit stepped away. In the weekly era, a handover like that would have been a natural moment for filler: a new studio finding its feet, a production pipeline being rebuilt, a schedule slipping. It is exactly the kind of disruption that recap episodes and filler arcs exist to absorb. The final season has no filler episodes at all.

That is worth pausing on, because it shows that the zero was not the achievement of one studio's good habits. It was the achievement of the structure: a finished manga, a seasonal schedule, and a production plan that treated the adaptation as a finite project with an ending. The studio changed and the structure held, and the structure is what kept the count at zero. It is also, incidentally, a useful answer to a question fans ask about every handover: whether the new studio will “ruin” the show. In this case the numbers say the handover changed the look of the show — the animation style is visibly different — and changed nothing about its fidelity. The classification does not care who animated an episode. It cares whether the episode adapts the manga, and all 89 of them do, or mostly do.

The final season's split into parts is the same story from the other side. A weekly show that falls behind schedule invents a filler arc. A seasonal show that needs more time splits its final season and asks the audience to wait. Attack on Titan's final season aired in parts across 2020, 2022 and 2023, and each part was awaited, argued over and watched in huge numbers — the waiting became part of the experience rather than a failure of it. The audience paid for zero filler in the currency the seasonal model charges: time. And, by every measure of attention the show received, they considered it a fair price.

The six mixed episodes

There is one judgement call in the whole run, and it is a small one: six mixed episodes, in our classification, out of 89. Mixed means part of the episode adapts the manga and part is original material — a canon scene padded with an invented subplot, or an original episode that ends with a genuine story beat. Six episodes in 89 is a low count by any standard — compare Naruto's 56, Shippuden's 64, or even Demon Slayer's 2 in 63, which is lower in absolute terms but a similar share — and it tells the same story as the zero filler count: this production almost never padded.

Our standing advice applies here as everywhere else: watch mixed episodes. The rule costs almost nothing on this series — six episodes in 89 — and it protects the canon content inside them, which in Attack on Titan's case is part of the main story's progression. The mixed-episode rule exists for series like Naruto, where 56 mixed episodes carry real story inside their padding; in Attack on Titan it is a formality, six episodes to watch rather than a judgement to make. But the formality is worth stating, because it is the same rule, applied consistently, and consistency is what makes a list trustworthy: the categories do not change their meaning from series to series, even when the counts do.

It is also worth saying what the six mixed episodes are not. They are not a hidden filler arc, and they are not a warning sign. A mixed episode is not a detour; it is a canon episode with some original material in it, and the canon material is the reason to watch. In a series with 83 canon episodes and 6 mixed ones, the practical advice is the simplest on the site alongside Demon Slayer's and Yu Yu Hakusho's: watch everything. The classification exists to tell you that, and to tell you why — not to give you something to skip.

What the waiting cost

There is a second half to this story, and it is the reason the zero-filler record took ten years to complete. A faithful adaptation that refuses to invent material can only move as fast as its source and its production allow — and Attack on Titan moved slowly. The first season aired in 2013. The last episode aired in November 2023. Ten years, four seasons, two studios, and long silences in between. The audience that started watching in 2013 waited four years for a second season, and waited again between the parts of the final season. The waiting was not an accident. It was the price of the plan, and it was charged in full.

That is the real trade-off, and it is worth stating plainly. The weekly model paid for its filler with completeness of a different kind: Naruto and Bleach ran for hundreds of episodes and reached their endings, filler and all, on a schedule the audience could set a watch by. The seasonal model pays for its purity with waiting: Attack on Titan is 89 episodes of superb adaptation that took a decade to finish, and its audience paid in years what a weekly audience pays in filler episodes. Demon Slayer's audience pays the same way, in gaps between seasons, and Hunter × Hunter's audience pays differently again — in absence, because that series stopped when its manga ran out and has not returned. There is no version of this story in which the audience gets everything: a complete adaptation, no filler, and no waiting. Every production chooses which shortage to accept, and Attack on Titan chose waiting.

None of that diminishes the achievement. It frames it. The 89 episodes that exist are among the most efficient long-form storytelling in anime, and their efficiency is measurable: 83 canon episodes out of 89, zero filler, six mixed episodes that you watch anyway. The cost is real, but it is a cost paid between seasons, not during them — and for the viewer deciding what to watch tonight, the episodes themselves are the thing. A new viewer in 2026 does not have to wait ten years. They have to watch 89 episodes, about 34 hours, and every one of them is the story.

How to watch it

The viewing instructions for Attack on Titan are the simplest on this site, and they fit in one sentence: watch episodes 1 to 89, all of them. That is the whole route. At two episodes a day, the 89-episode run takes about six and a half weeks; at one a day, about three months. There are no blocks to step over, no arcs to sample later, no route choices to make before episode one. The ranges table on the Attack on Titan series page confirms it: canon throughout, with six mixed episodes to watch as they come.

Two small pieces of advice for first-time viewers. First, trust the seasons. The show's long silences between seasons are part of its design, and the story is built to be re-entered: each season restates what you need, and the arcs — the fall of the walls, the coup, the basement, the war — are self-contained enough to survive a gap. Second, do not confuse a slow canon stretch with a detour. Attack on Titan's middle seasons spend long episodes on politics, history and conversation, and a viewer used to weekly filler may mistake that slowness for padding. It is not padding. It is the manga's own rhythm, adapted faithfully, and it is the reason the revelations land as hard as they do. A viewer who starts skipping will miss the story they came for.

Attack on Titan is the proof that the filler problem was never inevitable. Given a manga that finished first, a seasonal schedule with room to wait, and a production willing to stop between seasons rather than invent, a long-running anime can adapt its source episode for episode, with six mixed episodes as the only concession to the realities of television. Zero episodes out of 89. It is the smallest demand any series on this site makes of its viewers — matched only by Demon Slayer's zero in 63 and Yu Yu Hakusho's zero in 112 — and, for that reason, the easiest recommendation we can give: watch it all, skip nothing, and spend the time you would have spent skipping on the next series.

Filler lists in this article

Attack on Titan

89 episodes · Completed
0% filler

Demon Slayer: Kimetsu no Yaiba

63 episodes · Completed (TV)
0% filler

Hunter × Hunter (2011)

148 episodes · Completed
1.4% filler

More from the blog

← All articles