Resolution criteria

This market will resolve to the company that developed the #1 ranked model on the official Artificial Analysis LLM Leaderboard as of September 30, 2026, at 11:59 PM UTC.

• Primary Source: The Artificial Analysis LLM Leaderboard.

• Ranking Metric: The ranking will be determined by the Artificial Analysis Intelligence Index (e.g., v4.1.1 or any updated version active at the time of resolution).

• Tie-breaker: In the event that multiple models from different companies share the exact same top Intelligence Index score, the model displayed higher on the leaderboard UI will be deemed the winner. If no UI ordering distinguishes them, the market will resolve to a split among the tied companies.

• Model Ownership: The winning option will be the company that is the primary developer of the #1 model (e.g., "Claude" models resolve to Anthropic, "GPT" models to OpenAI, "Gemini" to Google, "Llama" to Meta, "Grok" to xAI, etc.).

• Other / New Answers: As the answer addition mode is set to "ANYONE", users can add other companies to the selection. If the #1 model belongs to a company not explicitly listed at the time of resolution, the market will resolve to "Other" (automatically added by Manifold).

Background

Artificial Analysis is an independent benchmarking platform that evaluates and ranks frontier AI models. Its primary metric for comparing overall capabilities is the Artificial Analysis Intelligence Index, a composite benchmark aggregating several evaluations across agentic tasks, coding, mathematics, and scientific reasoning.

As of late August 2026, the leaderboard is topped by Anthropic's Claude Opus 5 (with an Intelligence Index score of 63), followed by other leading models from OpenAI, Google, and open-weights releases from Alibaba (Qwen) and DeepSeek. This market tracks which AI laboratory's model will hold the top position at the end of September 2026.

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.