What independent estimates (OSINT, IISS, UK Ministry of Defence) say about Ukrainian killed and wounded from 2022 to 2025?

1. What the UN/HRMMU counts and does not count: verified civilian casualties, not military totals

The United Nations’ human‑rights monitoring mission (HRMMU/OHCHR) publishes cumulative, verified civilian casualties and has documented at least 14,534 civilian deaths and 34,115–38,472 civilian injuries across the conflict up to mid/late‑2025 in different UN summaries ^{[1] [5]}. These figures are explicit about methodology and severe under‑counting risk — they are not estimates of military killed or wounded and therefore do not provide a full picture of Ukrainian military casualties ^{[1] [5]}.

2. OSINT and investigative projects: detailed, public databases with limits

OSINT groups (e.g., Oryx, Mediazona/BBC collaborations, named‑list projects) document visually or name‑verified losses, mainly for equipment and for identified individual deaths; these sources produce robust minimums but explicitly undercount total casualties because many deaths are not publicly identifiable ^{[6] [7] [8]}. Media and investigative projects have been used as inputs by meta‑analyses (The Economist’s GitHub project collects and models many such sources), but they produce ranges rather than a single accepted tally for Ukrainian killed and wounded ^[9].

3. IISS and large institutional estimates: high Russian focus, Ukrainian implications

The International Institute for Strategic Studies (IISS) reports large manpower loss estimates for Russia (e.g., a January 2025 IISS breakdown that implied hundreds of thousands of Russian killed and wounded), and IISS assessments are often cited when analysts model both sides’ losses. IISS’s published figures focus more on Russia’s killed (~172,000) and wounded (~611,000) to early 2025 and are used by commentators to infer comparative Ukrainian losses; explicit IISS tallies for Ukrainian military killed/wounded are less prominent in the cited extracts ^[4].

4. CSIS, The Economist and “meta‑estimates”: combining sources produces wide ranges

Aggregators and policy centers (CSIS, The Economist) fit multiple data streams and intelligence fragments into statistical models. CSIS’s June 2025 analysis cited an estimate of roughly 400,000 Ukrainian killed or injured since 2022, including perhaps 60,000–100,000 deaths — but that estimate sits alongside other large, divergent tallies and is sensitive to model assumptions and input choices ^{[2] [9]}. The Economist published meta‑estimates that produced broad bands (hundreds of thousands killed and wounded across both sides) and made their data and code public to show sensitivity to inputs ^{[9] [10]}.

5. UK Ministry of Defence: frequent aggregate updates, focuses more on Russian totals

UK Defence Intelligence updates regularly report aggregate casualties and have repeatedly highlighted Russian killed‑and‑wounded totals (e.g., assessments of 900,000–1,118,000 Russian killed/wounded through 2025), while noting large annual additions in 2025; UK MoD public notes often reference Ukrainian tallies reported by Kyiv but do not publish a single independent Ukrainian military killed/wounded total in the material provided ^{[11] [12]}. The UK outputs are used in meta‑estimates but are primarily framed around Russian losses in public updates ^{[11] [12]}.

6. Ukrainian official tallies vs. external estimates: deliberate opacity and divergence

Ukraine’s government statements have provided totals (e.g., President Zelensky’s mid‑Feb 2025 figure of ~46,000 Ukrainian soldiers killed and ~390,000 wounded), which are cited by ISW and other outlets; Western intelligence and academic estimates sometimes track or contest these numbers, producing both higher and lower ranges depending on method ^{[13] [14]}. Public sources warn that all sides withhold or selectively publish casualty data for operational and political reasons, so official counts are contested and incomplete ^{[14] [13]}.

7. Why numbers differ and what to trust: methodology matters

Differences arise from (a) scope (civilian vs military; killed vs wounded vs missing), (b) method (name‑by‑name OSINT minimums vs modelled meta‑estimates vs verified UN tallies), and (c) institutional incentives (governments limit disclosure; OSINT groups can only verify visible cases) ^{[6] [9] [1]}. Where reporting converges — for example, the UN’s verified civilian counts — the numbers are authoritative for that narrow category; for overall Ukrainian military killed and wounded, available reporting shows wide, model‑dependent ranges rather than a single agreed figure ^{[1] [2] [9]}.

8. Bottom line for readers

Available independent and open‑source reporting does not provide a single, undisputed total of Ukrainian killed and wounded from 2022–2025. For civilian deaths and injuries, rely on HRMMU/OHCHR verified counts (e.g., ~14,534 civilian deaths, tens of thousands injured by late 2025) ^[1]. For military killed and wounded, use range‑based meta‑estimates (e.g., CSIS ~400,000 Ukrainian casualties with 60k–100k killed; various aggregators and the Economist provide modelled bands) and treat any single number as provisional and model‑dependent ^{[2] [9]}. Available sources do not mention a universally accepted independent tally that definitively pins Ukrainian military killed and wounded across 2022–2025.