Primary sourcelmarena.ai
View sourceEvent ·Multi-source
Agent Arena 分析:头部智能体每生成 1 token 需读取约 93 个输入 token
LMArena 对 26K+ 会话数据的分析显示,Agent Arena 排名前 10 的智能体每生成 1 个输出 token 平均要处理约 93 个输入 token。这意味着缓存输入的定价直接影响智能体运行成本,各实验室正竞相降低缓存输入价格。文章以 Anthropic 的 Opus 5.5 和 Google DeepMind 的 Gemini 4 Argon 的真实会话为例,演示了缓存在实际运行中的工作方式。
Current read
LMArena 拆了 2.6 万条智能体会话数据,发现输入输出 token 比例高达 93:1,看完你就懂缓存定价为什么成了价格战焦点。
11 sources43° AI heatLast updated 2026/11/2 19:08:00
Evidence
11 sourcesRelated source 1rohanpaul_ai
View sourceRelated source 2宝玉
View sourceRelated source 3Nathan Lambert
View sourceRelated source 4Decoder
View sourceRelated source 5arXiv: Anthropic
View sourceRelated source 6IT之家
View sourceRelated source 7VTV Công nghệ
View sourceRelated source 8Tech in Asia
View sourceRelated source 9shao__meng
View sourceRelated source 10Blognone
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.