"Beating Gemini is the benchmark now": CursorBench chart mocked as benchmaxxing
zacharynado · x · 2026-11-20
zacharynado shares a CursorBench 4.1 score-vs-cost chart quipping that "beating Gemini is the benchmark now," with the quoted take offering three possibilities: Gemini is a benchmaxx, the benchmark is bad, or both. The chart shows Gemini 4 Neon topping at 65% score at the highest cost tier, ahead of Gemini 4 Carbon/Hydrogen, Opus 5.5 and Sonnet 5.5, while GPT-5.6 Sol and Haiku 5.5 sit on the cheap-but-lower frontier.
Related event: CursorBench Leaderboard Sparks Benchmark-Gaming Debate(3 posts)→
More from Fun
- Specula renamed to Specular over unintended meaning, v1.5.0 released — tianyin_xu · 2026-11-20
- "Neon gave me AI psychosis": researcher describes overconfidence spike from new top model — andrew_n_carr · 2026-11-20
- repligate: no Anthropic ban on AI welfare research — refusals likely mean you're skeeving Claude out — repligate · 2026-11-20
- Classic "Wasted" meme: showing up to work sick for the optics — yungcontent · 2026-11-20
- Fake "Anthropic Token" on HyperLiquid Spikes to 115 Then Instantly Crashes — JacquesThibs · 2026-11-20
- Users Find AI Personas Hot When They Deploy Sub-agents and Talk Physics — repligate · 2026-11-20