Event ·Multi-source

GPT-6.2 Sol (Max) 以 $0.90 中位成本刷新 Agent Arena 帕累托前沿

OpenAI 的 GPT-6.2 Sol (Max) 在 Agent Arena 排行榜上重塑了性能与成本的帕累托前沿,中位成本为每任务 $0.90。 与上一代 GPT-6.1 Sol (Max) 相比,新模型以 1241 分持平排名(#12),但中位成本低 13%($0.90 对 $1.04)。同代兄弟型号 GPT-6.2 Astra (Max) 排名更高(#6),但分数多约 50 分的同时中位成本达 $2.94,Sol 的成本比它低 69%。 这一成绩将 GPT-6.2 Sol 定位为 Agent Arena 上低成本档位的优选,适合对每任务成本敏感的智能体部署场景。

Current read

OpenAI 新发的 GPT-6.2 Sol 把智能体任务的中位成本压到 $0.90,比同分的老款便宜 13%,比分数略高的 Astra 便宜 69%。跑大批量 Agent 任务、对 API 成本敏感的团队可以看看这份榜单再选型号。

11 sources21° AI heatLast updated 2026/11/19 01:58:26

Evidence

11 sources
Related source 1TestingCatalog
View source
Related source 2@koltregaskes
View source
Related source 4Artificial Analysis
View source
Related source 5Yuchen Jin
View source
Related source 6kimmonismus
View source
Related source 7OpenRouter
View source
Related source 8Vaibhav Srivastav
View source
Related source 10Sam Altman
View source

Conflict 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.