The Nihonto Price Index
What Japanese swords actually sell for, computed from the documented sales in our own registry. Dealers publish prose, auction data hides behind paywalls; nobody publishes open medians. So we computed them, with every sample size printed next to its number.
Read this before the numbers
Every statistic on this page is computed from the 26 single-blade auction results in our sales registry, where each sale carries its named source and citation. That gives these numbers an unusual property for this market: you can check every one of them against the underlying records on one page.
It also gives them limits, and we would rather state them than have you discover them. The registry is a curated record of notable public sales, which skews toward headline results; it is not a random sample of the sword market, where most blades trade privately in Japan for far less. Prices are nominal dollars in the year of sale, not adjusted for inflation. Where a house published a with-premium figure we use it; the 1986 and 1992 records do not state premium status, and we flag every place that matters. And the sample sizes are small: a median of four is a fact about four sales, not a law of nature. Each figure prints its n so you can weigh it yourself.
What NBTHK papers are worth
The lore says papers drive price. Our records say something more specific: the premium is modest between the lower tiers and enormous at the top. Median realized prices by the highest NBTHK designation stated in the sale catalog:
The honest reading: below Tokubetsu Juyo, what the blade is matters more than what the paper says. A great Juyo (the Naotsuna at $134,500) outsells a modest one (the Magoroku at $21,760) by six times inside the same tier. Papers narrow the argument; they do not settle it.
What the market pays for Kamakura
Sword lore holds the Kamakura period as the golden age. The modern-market records (2001 and later, so the 1990s bubble does not distort the comparison) put a number on the reverence:
One name complicates the Muromachi story in the right way: the registry's most famous Muromachi smith, Muramasa, sold for $35,840 in 2026, the top of the bucket. Legend is a pricing factor the period medians cannot see.
The one blade that sold twice
The registry contains exactly one documented repeat sale, and it is worth more than any average: the same physical object, priced by the same market, nine years apart. A Ko-Bizen tachi by Masazane, Heian period, Juyo Token no. 4117, exhibited at the Metropolitan Museum of Art in 2014, sold at Bonhams New York for $47,500 in October 2013 and again for $44,475 in October 2022.
One pair is one pair, and we will not build a law on it. But it is a documented counterweight to the "swords only go up" pitch, and it explains why our registry's patterns section treats investment framing with suspicion. A museum-exhibited, papered, thousand-year-old blade held for nine years returned less money than it cost, before expenses.
Why dealer prices and auction prices are different animals
The strongest blades in Japan mostly trade through dealers, not auctions, and dealer prices are asking prices: opening positions, not transactions. The registry's two dealer-published records make the gap vivid. A Tokyo dealer's 2025 asking price for a Juyo Kotetsu katana, JPY 60,000,000 (roughly $420,000), exceeds every realized auction price in our records since 1992. The same dealer's Ko-Ichimonji tachi at JPY 13,500,000 (roughly $94,000) sits right at our modern tachi median of $97,250 (n=8).
Neither number is dishonest; they are different instruments. An auction price is what someone actually paid on a stated day. A dealer price is what a specialist believes the right buyer will eventually pay, with the dealer's guarantee, patience, and reputation priced in. Comparing them directly is the most common mistake in sword-market conversation, and the two kinds of records are shaded differently in the registry for exactly that reason.
Every segment, every n
| Segment | n | Median | Range |
|---|---|---|---|
| All auction records, 1986-2026 | 26 | $96,000 | $15,625 to $418,000 |
| Modern era only, 2001-2026 | 20 | $45,988 | $15,625 to $238,000 |
| Record era, 1986-1992 | 6 | $143,000 | $104,500 to $418,000 |
| Tokubetsu Juyo Token | 4 | $187,750 | $175,000 to $238,000 |
| Juyo Token | 6 | $45,988 | $21,760 to $134,500 |
| Tokubetsu Hozon | 3 | $31,562 | $15,625 to $87,500 |
| Hozon | 1 | $20,075 | single record |
| Kamakura period (2001+) | 7 | $175,000 | $20,075 to $238,000 |
| Nanbokucho period (2001+) | 3 | $113,775 | $40,075 to $134,500 |
| Edo period (2001+) | 2 | $51,562 | $15,625 to $87,500 |
| Muromachi period (2001+) | 3 | $21,760 | $21,000 to $35,840 |
| Heian period (both sales of one blade) | 2 | $45,988 | $44,475 to $47,500 |
| Momoyama period | 1 | $27,500 | single record |
| Shinshinto and Meiji, 19th century | 2 | $39,781 | $31,562 to $48,000 |
| Tachi (2001+) | 8 | $97,250 | $20,075 to $238,000 |
| Katana (2001+) | 9 | $35,000 | $15,625 to $175,000 |
All figures Okami computed from the registry records; nominal USD; with-premium where the house stated it; GBP and EUR results use the approximate USD values published in the registry. The tachi-katana gap partly restates the period gap: most tachi in the records are Kamakura blades.
How this index is computed
- Source. The 26 single-blade auction results in our sales registry, each with a named source there. Nothing on this page rests on a record you cannot see cited.
- Medians, not averages. With samples this small and a $418,000 record in the set, an average would be a headline generator. Medians resist the outliers.
- Classification is ours. Period buckets and designation tiers follow the sale catalogs' own descriptions; the Go Yoshihiro tachi, catalogued late Kamakura to early Nanbokucho, is bucketed Kamakura, and the two-blade Shinkai daisho is one record. The classification table lives in our build script and every choice is reviewable.
- What we refuse to compute. No inflation-adjusted "real returns" (premium treatment is inconsistent before 2001), no per-smith values from single sales, no trend lines through 26 points. When the registry grows, the index recomputes; the number of records is part of every figure until then.
- Update cadence. Recomputed whenever the registry is updated, on its quarterly schedule after the spring and autumn auction seasons. Corrections with a source: info@okamiblades.com.
Last computed August 20, 2026, from the registry's 29 entries (26 single-blade auction results, 2 dealer-published asking records, 1 sale-event total). Okami Blades is a kitchen knife maker: we do not sell, appraise, or broker swords, and nothing here is valuation advice.
Related reading
- The Nihonto Sales Registry: every record behind these numbers, with citations.
- Blade auction records and the most famous knives in history.
- The Japanese sword section: what these objects actually are.
- The Smith Lineage Tree: several of the schools in these records, mapped.
