fun: 3/4 months ago I ran o3 for some academics on a set of AIME-style problems. It has taken them so long to write a summary of the results (96% irrc) that Alex solved proof & IMO in the meantime lol

fun: 3/4 months ago I ran o3 for some academics on a set of AIME-style problems. It has taken them so long to write a summary of results (98%) that Alex solved proof+IMO in the meantime lol

fun: 3/4 months ago I ran o3 for some academics on a private test set of AIME-like problems. It has taken them so long to write a summary of the results (96%) that Alex solved IMO in the meantime.

@alexwei_ can you tell the world how many tokens per problem?

To summarize this week: - we released general purpose computer using agent - got beaten by a single human in atcoder heuristics competition - solved 5/6 new IMO problems with natural language proofs All of those are based on the same single reinforcement learning system

17 M U.S. teens grades 9-12, ~5 US IMO golds in practice but ~20 kids at gold-level. So IMO gold is one-in-a-million math talent (for 18 year olds; but I bet next Putnam falls too). 99.9999th percentile.

Why am I excited about IMO results we just published: - we did very little IMO-specific work, we just keep training general models - all natural language proofs - no evaluation harness We needed a new research breakthrough and @alexwei_ and team delivered

If you claim to have seen this coming, you better have the manifold P&L to back it up

LLMs remain underrated

I feel this may be helpful to some of you today: