9h ago

Epoch AI finds evidence of acceleration on three of four capability metrics, driven by reasoning models like OpenAI o1-preview that now solve previously unresolved math problems

Reasoning models show steeper gains than non-reasoning models on math tasks.

——0——
Original post
OPEthan Mollick#175Ethan Mollick@EMOLLICK

Its funny how much the whole "strawberry" thing, which turned out to be o1-preview, was dismissed as overhyped at launch when it is clear in retrospect that it was way underhyped. A direct line from models unable to do basic math to solving unresolved math problems in 18 months.

9:52 PM · May 21, 2026 View on X
QUOTE POSTNick#411Nick@NICKCAMMARATA

@davidad actually you commented this that day on that post. I guess I updated a lot new scaling law, separate of results however good, do you think you sufficiently updated on new curve?

11:13 AM · May 22, 2026 · 1K Views
QUOTE POSTdavidad 🎇#458davidad 🎇@DAVIDAD

Yeah, this is what Ilya (fore)saw

Ethan MollickEthan Mollick@emollick

Its funny how much the whole "strawberry" thing, which turned out to be o1-preview, was dismissed as overhyped at launch when it is clear in retrospect that it was way underhyped. A direct line from models unable to do basic math to solving unresolved math problems in 18 months.

4:52 AM · May 22, 2026 · 49K Views
11:04 AM · May 22, 2026 · 20.4K Views
REPLYdavidad 🎇#458davidad 🎇@DAVIDAD

image from: https://epoch.ai/blog/have-ai-capabilities-accelerated

davidad 🎇davidad 🎇@davidad

Yeah, this is what Ilya (fore)saw

11:04 AM · May 22, 2026 · 20.4K Views
11:04 AM · May 22, 2026 · 953 Views
QUOTE POSTdavidad 🎇#458davidad 🎇@DAVIDAD

@nickcammarata Definitely not. I failed to update until I read the DeepSeek-R1-Zero paper

12:24 PM · May 22, 2026 · 104 Views