Be careful, this model is worse than 03-25 in 10 of the 12 benchmarks (!) I bet ... | Hacker News

Hacker Newsnew | past | comments | ask | show | jobs | submit

		arnaudsm 88 days ago \| parent \| context \| favorite \| on: Gemini 2.5 Pro Preview Be careful, this model is worse than 03-25 in 10 of the 12 benchmarks (!) I bet they kept training on coding, made everything worse on the way, and tried to hide it under the rug because of the sunk costs.

ramblerman 88 days ago | [–]

where do you see that?

arnaudsm 88 days ago | | [–]

New model homepage : https://deepmind.google/technologies/gemini/

Old model card : https://storage.googleapis.com/model-cards/documents/gemini-...

They intentionally buried that information

jstummbillig 88 days ago | [–]

It seems that trying to build llms is the definition of accepting sunk cost.

Consider applying for YC's Fall 2025 batch! Applications are open till Aug 4
Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact