METR 数据推算:AI 任务时长每 2 个月翻倍,Model 4 达 240 小时
scaling01 · x · 2026-11-07
METR published an analysis showing that since Oct 2025, AECI benchmark scores for Anthropic's models have grown at 1.9x [95%: 1.5–2.8] the earlier rate, while cautioning that how benchmark scores map to real-world capabilities remains unclear.
Extracting AECI scores from METR's figures, one analysis implies the task time horizon doubling time is now around 2 months, with Model 4's horizon at roughly 240 hours. The author stresses this is weak evidence — a clear 1:1 mapping from AECI to time horizon is uncertain at these magnitudes. The accompanying chart shows fitted doubling times accelerating from 260 days (2019–2023) to 92 days (Sonnet 4.5 era) to 56 days from Opus 4.5 onward.
「漫话AGI」频道最新
- 字节实验显示 Full Bandwidth Transformer 或以一半规模对等两倍大模型 — DimitrisPapail · 2026-11-07
- 开发者发帖求证:真有设计师因 AI 找不到工作而转行吗 — oykun · 2026-11-07
- 图灵奖得主 Patterson 断言：十年后经济产出将是现在的百倍 — davidpattersonx · 2026-11-07
- 研究员吐槽「非公开模型」造势游戏：厂商和评测者各赚声望 — ludwigABAP · 2026-11-07
- Every 创始人 Thesis 大会论点：模型普及后，护城河是你的判断力 — danshipper · 2026-11-07
- 前 OpenAI 政策高管:安全报告若不翻译成人话,传播效果将递减 — deanwball · 2026-11-07