Will GPT-5.3-Codex have a longer time-horizon than Claude Opus 4.6?
5
Ṁ100Ṁ127Mar 8
81%
chance
1H
6H
1D
1W
1M
ALL
Background
Both models were released on the same day. Both are their respective companies' frontier reasoning models. This is as close to a fair fight as we could possibly hope for. So, who is ahead? OpenAI or Anthropic?
Resolution Criteria
Resolves according to the score shown for each model on https://metr.org/blog/2025-03-19-measuring-ai-ability-to-complete-long-tasks/ (50% time horizon)
This question is managed and resolved by Manifold.
Market context
Get
1,000 to start trading!