ModelsMulti-source

OpenAI 发布 GPT-6.2 Astra:排名第 2 的前沿模型,价格降低 20%-600%

[AINews] GPT-6.2 Astra, the #2 Frontier Model is better and 20%-600% cheaper, but has mixed reviews; 6.2 Sol likely main workhorse

Why we picked it

OpenAI 一次放出了 GPT-6.2 Astra 和 Sol 两个模型,Astra 排第 2 但价格最多便宜 6 倍,测试者口碑却两极分化。做 API 成本优化的开发者可以看看定价表,再决定是切换还是继续观望。

2026 年 11 月 17 日,OpenAI 发布 GPT-6.2 系列的两个模型:Astra 和 Sol。Astra 在基准排名中位列第 2,价格比前代低 20%-600%。同期发布的 Sol 定位为日常主力工作模型,承担大多数常规任务。

Image · Latent Space (swyx)
Body

Astra 的定价调整幅度较大,最低降幅 20%、最高降幅 600%,意味着某些场景下的调用成本降到原来的几分之一。不过早期测试者对 Astra 的实际使用体验评价不一,模型在榜单成绩和真实工作流中的表现存在落差。Sol 则被定位为性价比主力,适合承接高频、常规的推理请求。

对开发者而言,这次发布的关键在于成本结构的变化:旗舰级能力以更低价格开放,可能改变复杂任务和轻量任务之间的模型选择策略。早期用户反馈的分歧也提示,榜单排名之外仍需自行实测后再决定切换。

(注:本条原始新闻正文极短,以上内容基于标题信息整理,未添加原文之外的数字。)

Original · Latent Space (swyx)

[AINews] GPT-6.2 Astra, the #2 Frontier Model is better and 20%-600% cheaper, but has mixed reviews; 6.2 Sol likely main workhorse

November 17, 2026 — GPT-6.2 Astra and Sol lower benchmark costs. Early testers disagree on what Astra is like to work with.

Industry Coordinates & Source Verification

This report is categorized under AITOP's "Models" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to US-China foundation model benchmarks and commercialization comparison and AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include GPT-6.2, OpenAI, Astra, reasoning models.

Source attribution: curated from Latent Space (swyx) (@swyx .ai). Cross-source corroboration: 11 independent sources have covered this development (including Artificial Analysis, Epoch AI, marktechpost), confirming industry-wide signal strength.