OpenAI发布GPT-6.2 Astra与Sol双模型
Story overview
OpenAI 发布 GPT-6.2 Astra 和 GPT-6.2 Sol 两款闭源模型,分别面向复杂专业工作与日常任务,已上线 API(gpt-6.2-astra、gpt-6.2-sol),并可在 ChatGPT Work 和 Codex 中使用;Plus、Pro 用户可直接使用,Enterprise 和 Edu 用户需由工作区管理员开启。 OpenAI 称 Astra 在 Agents' Last Exam 上比 GPT-6 Astra 高 5.2 分,成本不到其三分之一;量子位报道称 Astra 以 1318 Elo 登顶 AA-3D 榜单,Sol 编程能力接近 Astra,任务成本约为其五分之一。 API 定价上,Astra 每百万输入 token 8 美元、缓存输入 0.4 美元、输出 40 美元,输入与输出均较 GPT-6 Astra 下调 20%;Sol 沿用 GPT-6.1 Sol 的每百万输入 2 美元、输出 10 美元。输入超过 272K token 的请求,输入与缓存费率翻倍、输出按 1.5 倍计费;Ultrafast 在 2027 年 2 月 28 日前按标准价 3 倍收费,之后为 6 倍。
AI-generated from reports · Updated 14h ago
Report timeline
Follow the reports to explore different perspectives.
- 量子位 QbitAISelectedOpenAI releases GPT-6.2 Astra and Sol, dual flagship and lightweight models
OpenAI launched GPT-6.2 Astra, its flagship for hard projects, and GPT-6.2 Sol for frequent everyday tasks. Astra gains 5.2 points over its predecessor on Agents' Last Exam at under a third of the task cost, and tops the AA-3D leaderboard at 1318 Elo. Sol nears Astra on coding at about a fifth of the cost. Both are on ChatGPT Work, Codex and the API.
- SiliconANGLE · AISelectedOpenAI launches GPT-6.2 Astra for professional work and GPT-6.2 Sol for everyday use
OpenAI released GPT-6.2 Astra, a frontier model for complex professional work, and an updated everyday model, GPT-6.2 Sol. OpenAI says Astra scored 5.2 points above GPT-6 Astra on the Agents' Last Exam benchmark at under a third of the cost, and beat Anthropic's Claude Fable 5.2 by 5.5 points at 5% of the cost. API pricing is $8 per million input tokens and $40 per million output tokens.
- MarkTechPostSelectedOpenAI releases GPT-6.2 Astra and Sol, cutting Astra API prices 20%
OpenAI released two closed models, GPT-6.2 Astra for hard long-horizon work and GPT-6.2 Sol for everyday tasks, both live in its API with a 1,050,000-token context window. Astra's input and output prices fall 20% to $8 and $40 per million tokens. Artificial Analysis scores Astra 62 on its Intelligence Index, behind Claude Fable 5.2 at 65.
Story activity trend
Not enough continuous observations to show a trend yet.