Primary sourcelmarena.ai
View sourceEvent ·Multi-source
LMArena 榜单更新:Gemini 4 Carbon 登顶,GPT-6.1 Terra 入榜
LMArena 本周更新多个榜单排名。Google DeepMind 的 Gemini 4 Carbon (High) 拿下 Text Arena 第一、Agent Arena 第二、Code Arena: WebDev 第四,在三个榜单上重塑了 Pareto 前沿。OpenAI 的 GPT-6.1 Terra (Max) 进入 Agent Arena 第八名,净提升 10.6%,每任务中位成本 0.62 美元,主打性价比。微软 MAI-Image-2.7-Preview 在 Single Image Edit Arena 排名第三,比上一版 2.6 的第六名高出 45 分。此外 LMArena 给 Agent Mode 新增了 Connectors,可在单个任务中组合 Vercel、Linear、GitHub 等工具完成复杂工作流。
Current read
三大模型同周换榜:Gemini 4 Carbon 直接冲到文本第一,GPT-6.1 Terra 花 0.62 美元就能跑完一个任务,选模型前先看看这份数据。
11 sources71° AI heatLast updated 2026/10/30 18:43:56
Evidence
11 sourcesRelated source 1爱范儿
View sourceRelated source 2rohanpaul_ai
View sourceRelated source 3宝玉
View sourceRelated source 4Simon Willison’s Weblog
View sourceRelated source 5小互
View sourceRelated source 6OpenAI Blog
View sourceRelated source 7Healthcare Dive
View sourceRelated source 8Decoder
View sourceRelated source 9The Straits Times: Business
View sourceRelated source 10Tech in Asia
View sourceConflict check
Dedup data only shows that these sources discuss the same event, not that they agree. There is no structured supporting or refuting evidence yet, so no conflict is inferred automatically.