What is the median advertised salary in Tokyo?
¥8.3M a year — the median midpoint of 3,729 complete advertised ranges; 57% of live jobs publish pay (as of October 8, 2026).
Median advertised pay in Tokyo, broken down by category and skill, and how it differs alongside posting language and visa signals.
¥8.3M a year — the median midpoint of 3,729 complete advertised ranges; 57% of live jobs publish pay (as of October 8, 2026).
Software & IT, with a median advertised salary of ¥9.0M across 1,340 salary ranges (as of October 8, 2026).
Jobs asking for Kubernetes have the highest median advertised salary, ¥10.0M (105 salary ranges). This is an association, not proof the skill causes the difference (as of October 8, 2026).
| Job category | Live jobs | Median salary | Salary sample |
|---|---|---|---|
| Business & Corporate Roles | 3,201 | ¥8.0M | 1,599 ranges |
| Software & IT | 2,416 | ¥9.0M | 1,340 ranges |
| Creative & Media | 571 | ¥7.2M | 355 ranges |
| Engineering & Infrastructure (Physical) | 506 | ¥8.0M | 303 ranges |
| Customer Support & Service | 213 | ¥6.5M | 98 ranges |
| Other | 61 | ¥5.8M | 27 ranges |
| Education & Teaching | 21 | ¥6.5M | 7 ranges |
| Skill | Live jobs | Median salary | Salary sample | vs market |
|---|---|---|---|---|
| Kubernetes | 178 | ¥10.0M | 105 ranges | +¥1.8M vs market |
| Next.js | 126 | ¥10.0M | 85 ranges | +¥1.8M vs market |
| Kotlin | 89 | ¥10.0M | 55 ranges | +¥1.8M vs market |
| Terraform | 284 | ¥9.8M | 175 ranges | +¥1.6M vs market |
| Go | 198 | ¥9.8M | 120 ranges | +¥1.5M vs market |
| Generative AI / LLM | 719 | ¥9.5M | 414 ranges | +¥1.3M vs market |
| Google Cloud | 426 | ¥9.5M | 249 ranges | +¥1.3M vs market |
| TypeScript | 355 | ¥9.5M | 215 ranges | +¥1.3M vs market |
| PostgreSQL | 146 | ¥9.5M | 104 ranges | +¥1.3M vs market |
| Node.js | 99 | ¥9.5M | 63 ranges | +¥1.3M vs market |
| Signal | Live jobs | Median salary | Published ranges | Coverage |
|---|---|---|---|---|
| Japanese posting | 5,857 | ¥8.3M | 3,608 | 62% |
| English posting | 775 | ¥10.5M | 39 | 5% |
| Mixed-language posting | 357 | ¥9.6M | 82 | 23% |
| Signal | Live jobs | Median salary | Published ranges | Coverage |
|---|---|---|---|---|
| Visa maybe | 2,405 | ¥9.3M | 951 | 40% |
| No visa signal | 4,547 | ¥8.0M | 2,773 | 61% |
| Signal | Live jobs | Median salary | Published ranges | Coverage |
|---|---|---|---|---|
| JLPT N2 | 130 | ¥8.0M | 56 | 43% |
| JLPT N1 | 188 | ¥9.3M | 68 | 36% |
| Native Japanese | 201 | ¥10.0M | 62 | 31% |
Annual base salary in millions of yen. Each job is represented by the midpoint of its published range.
Skills ranked by median salary, with at least 8 complete published salary ranges. Association does not prove the skill caused the difference.
How advertised salary differs alongside posting language and estimated visa support. These are unadjusted associations—not evidence that language ability or sponsorship caused the salary difference.
Only groups with at least 20 complete salary ranges are published. Coverage can differ sharply because many employers do not advertise pay, and the mix of occupations, seniority and employers differs between groups.
Posting language is observable, but it is not the same as a spoken-language requirement. English postings currently publish salary much less often, so their median may not represent the wider English-friendly market.
Visa likely means the listing has a strong AI-extracted or corroborated sponsorship signal; Visa maybe is weaker evidence. Neither is an employer guarantee, and “No visa signal” does not mean sponsorship is unavailable.
How to read these numbers responsibly.
6,989 live employer job pages explicitly located in Tokyo or a named Tokyo ward/neighborhood. Duplicate IDs and expired listings are removed; 2,216 Japan-wide or non-Tokyo listings are excluded.
A skill counts when the posting's structured vector is at least 70% confident. This includes explicit requests and strong deterministic implications such as React implying JavaScript.
Only valid ranges with both a minimum and maximum are used. We take each range midpoint, then calculate the median or mean across jobs. The complete-range coverage is 53%.
The current snapshot refreshes after the job crawl. Daily aggregate snapshots are retained, and trend comparisons appear after 14 comparable days. Skill, language and visa-support fields are AI-assisted estimates; verify details on the employer's original posting.