Compassion Benchmark
Independent benchmark research · no paid rankings

Score how 1,256 institutions recognize and reduce suffering

Governments, companies, AI labs, and cities — re-examined every weekday, sourced from public evidence, free.

Institutional compassion is how reliably an institution recognizes, responds to, and reduces the suffering of the people it affects. The benchmark makes this measurable, comparable, and difficult to fake.

No entity pays for inclusion, score changes, or suppression of findings. About the institution →

2026-07-12 Anthropic holds at 59.1 after a Chinese regulator's backdoor allegation fails both tests for a score change.

Browse the 7 indexes →

1,256Entities currently ranked and scored
7Published index families
8Core benchmark dimensions
40Subdimensions in full standard

Current publication set

IndexCoverage
World Countries207 countries and territories
U.S. States50 states and the District of Columbia
Fortune 500447 published company rankings
AI Labs50 leading AI labs
Humanoid Robotics LabsTop 50 global labs
U.S. CitiesTop 150 American cities
Global CitiesTop 250 cities worldwide

Public rankings are independent. Commercial offerings support access and interpretation — never paid score changes.

Today's briefing · 2026-07-12

Anthropic holds at 59.1 after a Chinese regulator's backdoor allegation fails both tests for a score change.

1,260 scanned · 15 assessed · 0 changes — all confirmedFull briefing →
boundary watch

Anthropic Holds at 59.1 as a Chinese Regulator's Backdoor Claim Fails the Benchmark's Own Test

high

Anthropic holds a published score of 59.1 out of 100, in the Functional band. Between July 8 and 10, a cybersecurity platform run by China's Ministry of Industry and Information Te

documented

Tunisia's Pending Score Cut Is Reaffirmed for a Third Day as 21 Opposition Figures Are Sentenced

high

Tunisia holds a published score of 34.4 out of 100, in the Developing band. A review proposed on July 10 suggested a lower score of 23.8 out of 100, based on the 25-year prison sen

Lead signal score position

Anthropic Holds at 59.1 as a Chinese Regulator's Backdoor Claim Fails the Benchmark's Own Test

Ai Labs score position on the 0–100 band scale (Critical · Developing · Functional · Established · Exemplary)

Anthropic score: 59.1 — in the Functional band (40–60). 0.9 points to the Established band.59.10.9 pts to Established

Liked today's briefing? Get the Friday email — free.

Weekly score highlights — institutional compassion findings

The week's top score movements and evidence-linked findings across 1,256 entities, delivered every Friday. Daily briefings publish on the site. Free.

No spam. Unsubscribe anytime. Your email is never shared.

How the benchmark works

The framework measures institutions across eight dimensions: Awareness, Empathy, Action, Equity, Boundaries, Accountability, Systemic Thinking, and Integrity.

Published scores are derived from public evidence and normalized to a 0–100 scale. Consistency is rewarded: strong, even performance across all eight dimensions earns up to +10 points; any dimension at zero (active harm) cancels the bonus.

Why it's difficult to fake: Consistency is rewarded: strong, even performance across all eight dimensions earns up to +10 points; any dimension at zero (active harm) cancels the bonus. A single dimension at zero (documented active harm) cancels the consistency bonus, so surface-level improvements without addressing root failures do not improve the composite. Full scoring methodology →

View methodology

Current research program

The benchmark has moved from concept to active publication. The current program includes live index reports, a formal methodology, a broader 40-subdimension standard, and an operating model for future data products, certified assessments, and enterprise benchmark services.

Explore research

The state of institutional compassion

Of 1,256 entities benchmarked across governments, corporations, AI labs, and cities — most cluster in the middle. The chart below shows band distribution across the full catalog.

Flagship Annual Report — 2026

The State of Institutional Compassion — 2026

Across 1,256 institutions, the modal result (the band containing the most entities) is mediocrity — 67.7% cluster in the Functional band (40–60) on the chart below. A 90.5% equity gap persists at the bottom of every index family, meaning the best-scoring entities score nearly 90 points higher than the lowest-scoring ones.

Read the 2026 report →

Band distribution — all entities

1,256 institutions. Each segment shows the share in that band. Read left-to-right: Critical (0–20) to Exemplary (80–100).

17914%55945%32025%13511%Critical 0–20Developing 20–40Functional 40–60Established 60–80Exemplary 80–100
Source: Compassion Benchmark · CC-BY

Source: Compassion Benchmark · CC-BY

Cite this chart

Citation

Compassion Benchmark — Band distribution — all entities. compassionbenchmark.com/. CC-BY 4.0.

How to read the scores

The 0–100 scale — five bands

Every entity — state, corporation, AI lab, robotics lab, or city — is scored 0–100 across 8 dimensions and 40 subdimensions. The composite score places the entity in one of five bands:

Critical0–20Foundational compassion practices are absent or documented active harm is present.
Developing20–40Some practices are emerging but remain inconsistent, reactive, or unevenly applied.
Functional40–60Core practices exist and meet a basic bar, with significant gaps remaining.
Established60–80Practices are systematic, documented, and supported by consistent evidence.
Exemplary80–100Practices are independently verified, consistent, and sustained under pressure.

The 8 dimensions

Each dimension is scored 1–5 across 5 subdimensions (40 subdimensions total), then converted to a 0–100 composite. A score of 1.0 on a subdimension represents the minimum anchor; 5.0 is exemplary conduct.

AWRAwarenessDoes this entity reliably detect when others are in pain or need — before they name it?
EMPEmpathyDoes this entity genuinely connect with the inner experience of those it serves?
ACTActionDoes compassionate understanding translate into real, proportional, effective help?
EQUEquityIs care distributed fairly — especially toward those with greatest need and least power?
BNDBoundariesIs helping sustainable, ethical, and autonomy-preserving — not dependency-creating?
ACCAccountabilityDoes this entity own its failures, correct course, and make genuine repair?
SYSSystemic ThinkingDoes compassion extend to root causes and structural change — not only symptom relief?
INTIntegrityIs compassion genuine, consistent, and non-performative — especially when it costs something?

Scores are based on public evidence — government reports, regulatory filings, independent audits, judicial findings, and verifiable third-party records. Entities never pay for inclusion, score changes, or suppression of findings. Full methodology

The 8 dimensions of institutional compassion

Every entity is scored across the same 8 dimensions — from awareness of suffering to structural integrity. Together they define what it means to institutionally reduce harm.

The 8 dimensions

Framework diagram — representative shape, not a scored entity. Each axis scored 0–5.

Framework (representative) — compassion dimension profile radar12345Awareness: 3.5/5.0AWR3.5Empathy: 3.5/5.0EMP3.5Action: 3.5/5.0ACT3.5Equity: 3.5/5.0EQU3.5Boundaries: 3.5/5.0BND3.5Accountability: 3.5/5.0ACC3.5Systemic Thinking: 3.5/5.0SYS3.5Integrity: 3.5/5.0INT3.5Framework (representative)

Note: radar area can visually exaggerate differences — read the per-axis values, not the area.

Framework diagram — representative shape only, not a scored entity. Source: Compassion Benchmark · CC-BY

Source: Compassion Benchmark · CC-BY

Cite this chart

Citation

Compassion Benchmark — The 8 dimensions. compassionbenchmark.com/. CC-BY 4.0.

AWR Awareness ·EMP Empathy ·ACT Action ·EQU Equity ·BND Boundaries ·ACC Accountability ·SYS Systems ·INT Integrity

AWRAwareness

Does this entity reliably detect when others are in pain or need — before they name it?

EMPEmpathy

Does this entity genuinely connect with the inner experience of those it serves?

ACTAction

Does compassionate understanding translate into real, proportional, effective help?

EQUEquity

Is care distributed fairly — especially toward those with greatest need and least power?

BNDBoundaries

Is helping sustainable, ethical, and autonomy-preserving — not dependency-creating?

ACCAccountability

Does this entity own its failures, correct course, and make genuine repair?

SYSSystemic Thinking

Does compassion extend to root causes and structural change — not only symptom relief?

INTIntegrity

Is compassion genuine, consistent, and non-performative — especially when it costs something?

Full methodology → · Source: Compassion Benchmark · CC-BY

Seven indexes at a glance

How to read: each bar shows the share of entities in each band — Critical (red) → Developing → Functional → Established → Exemplary (blue). Sorted by share in the top two bands.

Shared band legend:CriticalDevelopingFunctionalEstablishedExemplary
Humanoid Robotics LabsView index →

Earliest-stage sector — most labs in Developing.

1020%1020%1530%1326%Critical 0–20Developing 20–40Functional 40–60Established 60–80Exemplary 80–100
Source: Compassion Benchmark · CC-BY

No lab yet reaches Established; safety gaps dominate.

510%1836%1530%816%4Critical 0–20Developing 20–40Functional 40–60Established 60–80Exemplary 80–100
Source: Compassion Benchmark · CC-BY
Global CitiesView index →

Cities cluster in Functional; equity drives variance.

6827%10642%3514%2510%Critical 0–20Developing 20–40Functional 40–60Established 60–80Exemplary 80–100
Source: Compassion Benchmark · CC-BY
U.S. CitiesView index →

U.S. cities span all five bands; regional spread is wide.

10271%3826%Critical 0–20Developing 20–40Functional 40–60Established 60–80Exemplary 80–100
Source: Compassion Benchmark · CC-BY
World CountriesView index →

Most countries in Developing; Critical band is large.

4724%7740%2915%2613%14Critical 0–20Developing 20–40Functional 40–60Established 60–80Exemplary 80–100
Source: Compassion Benchmark · CC-BY
Fortune 500View index →

Corporations cluster in Developing–Functional; top band rare.

5312%21648%11726%5512%Critical 0–20Developing 20–40Functional 40–60Established 60–80Exemplary 80–100
Source: Compassion Benchmark · CC-BY
U.S. StatesView index →

States span Developing through Established; no Exemplary.

419%943%838%Critical 0–20Developing 20–40Functional 40–60Established 60–80Exemplary 80–100
Source: Compassion Benchmark · CC-BY

Independence policy

  • Entities do not pay to be included in an index
  • Entities do not pay to improve their score or rank
  • Commercial services do not suppress public findings
  • Benchmark methodology is not sold to fit a preferred outcome
  • Paid offerings support access, interpretation, review, and institutional use

What is the Compassion Benchmark?

What is the Compassion Benchmark?
The Compassion Benchmark is an independent benchmark institution that measures how reliably institutions recognize, respond to, and reduce the suffering of the people they affect. It publishes comparative rankings across 1,256 entities — governments, corporations, AI labs, cities — re-examined every weekday, sourced from public evidence, free to access.
How are scores calculated?
Every entity is scored across 8 dimensions (Awareness, Empathy, Action, Equity, Boundaries, Accountability, Systemic Thinking, Integrity), each on a 0–5 scale, normalized to a 0–100 composite. Consistency is rewarded: Consistency is rewarded: strong, even performance across all eight dimensions earns up to +10 points; any dimension at zero (active harm) cancels the bonus. Scores are derived from public evidence only — not from self-reporting or paid submissions.
What do the bands mean — Critical, Developing, Functional, Established, Exemplary?
The five bands divide the 0–100 scale into equal 20-point segments. Critical (0–20): foundational practices absent or active harm documented. Developing (20–40): some practices emerging but inconsistent. Functional (40–60): core practices exist with significant gaps. Established (60–80): practices systematic, documented, evidenced. Exemplary (80–100): independently verified, consistent, sustained under pressure.
How often are scores updated?
The nightly pipeline scans the monitored entity catalog and publishes a daily briefing every weekday. Score changes are proposed when new public evidence crosses the scoring threshold for a dimension change; most days include confirmations (scores held at published values after re-examination) alongside any proposals. The daily briefing is published free on the site.

About the institution → · Today's briefing (2026-07-12) → · Methodology →

The benchmark updates every weekday. Come back tomorrow.

Each day's briefing covers which entities were scanned, which were assessed, and which crossed a scoring threshold — with the evidence. Free, on the site. Or get the Friday weekly highlights by email.