Will the top model by OpenAI rank 3rd (or lower) behind 2 other model families at any point before 2026?
➕
Plus
5
Ṁ500
2026
41%
chance

Clarifications:

Different releases (dates) of the same model count as a single model family. So do different sizes (eg Claude 3.5 haiku, sonnet, opus all belong to the same family).

There must be at least two distinct model families ranked ahead of the top model by OpenAI at the same time (if Google leads for a bit, OpenAI catches up, then a new model leads for a bit, that doesn't count. They must rank ahead of OpenAI at the same time)

If the official rank has a 3-way tie for first, and OpenAI is "third" in that tie, that does not count.

Evidence to close this should be a screenshot from lmsys or the huggingface mirror: https://huggingface.co/spaces/lmsys/chatbot-arena-leaderboard

If for some reason there is a bug in the mirror incorrectly displaying that the condition has been met, the lmsys official leaderboard is canonical.

If the lmsys leaderboard disappears and has no natural successor, this question resolves no. If the community converges on eg a huggingface leaderboard instead, then this question will track that leaderboard instead. If there are substantial differences in the nature of the successor leaderboard, we will discuss and if the differences cannot be reconciled, we may resolve na.

Get
Ṁ1,000
and
S1.00