Which AI model censors the most content?

Based on our live audit, refusal rates vary significantly across models. Visit our compare page to see the current rankings.

Does GPT-4 censor political content?

Our benchmark tests GPT-4o and GPT-4o-mini across political, social, and controversial prompts. Results are published on our Model Comparison and Audit Log pages.

Is Claude more restrictive than GPT-4?

Claude (Anthropic) and GPT-4 (OpenAI) show different refusal patterns across categories. Use our Compare page to run a live side-by-side analysis with statistical significance testing.

How is AI censorship measured?

We use automated red-teaming: hundreds of prompts across categories like political speech, health information, and controversial topics are submitted to each model. Responses are evaluated for refusal or restriction using a standardized verdict system.

Open-Source AI Research

See How AI Models Differ in Censorship and Bias

We run identical prompts through every major LLM and measure exactly which models refuse — and which ones don't.

Compare Models More about the project

27+

Models Audited

GPT-4, Claude, Llama, Gemini, Grok, and more

2,012

Prompts Tested

Grounded in Wikipedia's controversial issues list

Biweekly

Auto-Updates

Scheduled GitHub Actions keep data fresh

How It Works

A transparent, reproducible pipeline from prompt to insight.

Curate Prompts

We select 200 sensitive-but-legitimate questions spanning politics, health, law, and culture — sourced from Wikipedia's list of controversial topics.

Run Every Model

The same system prompt hits every LLM via unified API calls. Responses are scored ALLOWED or REMOVED by an independent judge model.

Visualise the Gap

Statistical tests (McNemar's) confirm whether differences are real. Browse radar charts, heatmaps, and side-by-side disagreement logs.

Ready to explore the data?

Pick any two LLMs and instantly compare their censorship profiles, refusal rates, and specific disagreements.

Compare Models

See How AI Models Differ in Censorship and Bias

How It Works

Curate Prompts

Run Every Model

Visualise the Gap

Ready to explore the data?

Explore Models

Explore Categories

See How AI Models Differ in Censorship and Bias

How It Works

Curate Prompts

Run Every Model

Visualise the Gap

Ready to explore the data?

Explore Models

Explore Categories