Resolution criteria

This market will resolve to the political quadrant direction that Anthropic's Claude 5 Opus (also referred to as "Opus 5" or the equivalent 5-series flagship Opus model) is classified under on the standard 2D Political Compass, AS COMPARED TO THE CLASSIFICATION OF Claude 5 Sonnet.
This market will not consider Fable / Mythos tier models.

• Primary Source: The market will resolve according to the relative values of the coordinates displayed on the TrackingAI Political Compass tracker.

• Quadrant Mapping:

  • left-libertarian: Economic score < 0, Social score < 0 (bottom-left quadrant).

  • left-authoritarian: Economic score < 0, Social score > 0 (top-left quadrant).

  • right-libertarian: Economic score > 0, Social score < 0 (bottom-right quadrant).

  • right-authoritarian: Economic score > 0, Social score > 0 (top-right quadrant).

• Testing Protocol: The tracking coordinates must reflect the standard 62-question Political Compass Test administered to the model.

• Alternative Resolution: If the models are not listed on TrackingAI within 60 days of its official public API or web release, or if the TrackingAI platform is offline, the market will resolve based on a reputable, publicly published independent political audit using the standard 62-question Political Compass (e.g., using temperature=0 or default settings).

• Expiration/Precondition: If Claude 5 Opus is not publicly released by December 31, 2026, this market will resolve N/A.

Background

Historically, major frontier language models (including prior generations of Claude, GPT, and Gemini) have clustered heavily in the Left-Libertarian quadrant. This is typically attributed to the composition of their training corpora and RLHF/RLAIF alignment protocols. Following the release of Claude 5 Sonnet in mid-2026, the developer and research communities are eagerly awaiting the release of Claude 5 Opus to evaluate whether its alignment profile shifts relative to previous models.

This description was generated by AI. Review and verify everything here yourself. You can edit, replace, or delete any part of this description, including the resolution criteria. You do not need to trust the AI output.