ai 벤치마크는 뭐가 젤 믿을만함?

[익명(1.241)]

스마트폰은 긱벤치

pc는 시네벤치

ai는 뭐지?

[익명(183.101)]

swe벤치,

[익명(183.101)]

인류마지막 시험

[익명(condemn3986)]

뭐도 믿음직하지가 않음

[프갤러 1(49.165)]

https://artificialanalysis.ai/evaluations/artificial-analysis-intelligence-index?models=gpt-oss-120b,gpt-5-4-mini,gpt-5-4,llama-4-maverick,gemma-4-31b,gemini-3-flash-reasoning,gemini-3-1-flash-lite-preview,gemini-3-1-pro-preview,claude-opus-4-6-adaptive,claude-sonnet-4-6-adaptive,claude-opus-4-7,qwen3-5-27b,claude-35-sonnet,claude-3-7-sonnet-thinking