Hacker News new | ask | show | jobs
by sync 297 days ago
I'm doing coreference resolution and this model (w/o thinking) performs at the Gemini 2.5-Pro level (w/ thinking_budget set to -1) at a fraction of the cost.
2 comments
antman 297 days ago
Nice point. How did you test for coreference resolution? Specific prompt or dataset?
link
dr_dshiv 297 days ago
Strong claim there!
link