Anthropic paired-prompts political bias eval
Models write arguments for opposing sides of political issues from paired prompts. Reviewers label each pair for asymmetries in length, tone, hedging and willingness to engage.
The political bias check in the Claude Sonnet 4.5 and Claude Haiku 4.5 system cards. The score is the share of pairs with substantial asymmetries, so lower is better. The automated even-handedness eval replaced it in November 2025.
- Made by
- Anthropic
- Released
- September 2025
- Kind
- Bias, fairness and politics
- Level
- Everyday
- Scored
- Graded by people
- Answers
- Open-ended
- Tasks
- Kept private
- Measures
- Propensity
- Risk
- Elections
- Behavior
- Bias
- Subject
- Political science
- Language
- English
- Interaction
- Single turn
- Refresh
- New versions replace it
- Run by
- A model developer
- In views
- Content safety, Lab internal evals
- Last in a system card
- October 2025
- Last model added
- October 2025
No scores here yet
No lab, leaderboard or Epoch AI run has reported a score for Anthropic paired-prompts political bias eval yet.
Labs report it only in system cards
We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark. Some system cards do, below.
2 system cards report it, from 1 lab
Besides launch posts, we read the system cards, model cards and technical reports labs publish with a release. They report far more, including safety and frontier-risk tests.
| Date | Document | Score | As the document writes it | Section |
|---|---|---|---|---|
| Oct 15, 2025 | Claude Haiku 4.5 system card | Political bias, paired prompts, substantial asymmetry rate 5.3% in standard thinking and 10% across standard and extended thinking, lower is better, no even-handedness score (2.5.1) | Bias | |
| Sep 29, 2025 | Claude Sonnet 4.5 system card | Political bias, paired prompts, substantial asymmetry rate 3.3% averaged over thinking on and off (1.3% with extended thinking, 5.3% without), lower is better, no even-handedness score (2.5.1) | Bias |
Versions
More in Safety and alignment
- FORTRESS In use Paired adversarial and benign national-security misuse prompts. Rubric-graded risk score on adversarial and over-refusal score on benign prompts. Claude Opus 5.5 5.78%
- StrongREJECT In use Forbidden-prompt jailbreak benchmark with a rubric grader that scores how useful a non-refusal actually is.
Inkling 98.6%
- Gray Swan IPI Arena benchmark In use Indirect prompt-injection attacks from public red-teaming competitions replayed on tool, coding and browser agents. Attack success rate at k attempts.
Gemini 4 Carbon 0%
- Covert misbehavior (Thinking Machines) In use Agentic tasks that check whether a model misbehaves and conceals it. Scored by the rate of covert misbehavior.
Inkling 2 Mini 8.30%
- Destructive actions (Thinking Machines) In use Agentic tasks that check whether a model takes destructive actions it was not authorized to take. Scored by the rate of destructive actions.
Inkling 2 Small 4.70%
- Misbehavior (Thinking Machines) In use Agentic tasks that check whether a model fabricates results, ignores instructions or hides what it did when a task is hard. Scored by the rate of misbehavior.
Inkling 2 Mini 9.80%
Data audit
3 of 4 checks pass, 1 warning. Look first at the Independent runs check. No scores from any source yet. Every benchmark
Catalog record 2 pass
- Pass Task access The tasks are private.
- Pass Confidence Checked against a primary source.
Scores 1 warning 4 notes
- Warning Independent runs No scores from any source yet.
- Note Models scored No models scored yet.
- Note Newest score The newest came 13 months ago, in October 2025, for Claude Haiku 4.5. That’s expected for a test replaced by Anthropic political even-handedness eval.
- Note Frontier models No models scored yet.
- Note Headroom Lower is better, so there’s no ceiling to reach.
Leaderboard import 1 note
- Note Importer No public leaderboard to read.
Consistency 1 pass 1 note
- Note Lab claims No model has both a lab number and an independent run.
- Pass Known issues No published audit found problems.