OpenAI 推出面向金融服务的 ChatGPT，内置金融数据与 GPT-6 Astra
OpenAI 于 9 月 10 日发布「ChatGPT for Financial Services」，一款面向金融行业的定制化 ChatGPT Work 体验，将内置金融数据与前沿推理能力结合，帮助团队撰写研究报告、搭建财务模型并制作客户材料。
该产品由 OpenAI 与摩根士丹利（Morgan Stanley）、Evercore 共同设计打磨。合作让 OpenAI 精准定位金融机构面临的最大痛点，并据此打造能切实提升每位银行家日常效率的方案。
产品内置来自 Daloopa、PitchBook、LSEG News、Crunchbase 等供应商的优质金融数据，无需再配置 MCP 连接器。这些数据由 OpenAI 统一索引与托管，从而支持更精准的检索，以及可将数字与论断回溯到原始出处的细粒度引用（granular citations）。
在模型能力上，产品原生支持包括 GPT-6 Astra 在内的前沿智能，并会在新模型发布后开箱即用。企业可集中管理访问权限与数据连接，依托 ChatGPT 企业级安全与治理控制。
早期合作聚焦于投行业务与股票研究——可靠的数据访问与高质量产物生成是合作方团队的最大痛点。OpenAI 表示，与伙伴的协作将反哺后续训练、产品改进，并推动向其他金融服务品类的拓展。
数据供应方面，OpenAI 正与 S&P Capital IQ、LSEG、MSCI、Dow Jones Factiva、Moody's 推进统一登录与授权集成，用户通过 ChatGPT 登录即可自动获得已订阅数据的访问权限。
连接器生态已包含 50 余个，涵盖 Datasite、Box、Preqin、FactSet、Intapp 等；OpenAI 还重点优化了金融服务中最常用的 MCP，让团队少排错、多分析。
在安全合规上，产品基于 ChatGPT Enterprise 的 SAML SSO、SCIM 配置与基于角色的访问控制，支持信息隔离墙（information barriers）与多工作区；企业业务数据默认不用于训练模型，传输与存储均加密，合规团队可将工作区日志导出至既有审计流程。
管理员可通过专属后台发布 Excel、Word、PowerPoint 模板，使团队按公司格式与风格一键生成估值模型、研究笔记与推介材料（pitchbook）。
该产品面向符合条件的金融机构开放，有意者可通过账户团队联系 OpenAI。
本文来源：https://openai.com/index/introducing-chatgpt-financial-services/
若有版权问题，请来信告知：13951431913#139.com 或在右方提交评论。
OpenAI Unveils ChatGPT for Financial Services, Featuring Integrated Financial Data and GPT-6 Astra
On September 10, OpenAI introduced "ChatGPT for Financial Services," a specialized ChatGPT Work experience tailored for the financial industry. This solution combines integrated financial data with cutting-edge reasoning capabilities to assist teams in drafting research reports, building financial models, and creating client materials.
The product was developed in collaboration with Morgan Stanley and Evercore, allowing OpenAI to pinpoint the most significant pain points faced by financial institutions. The resulting solution is designed to visibly enhance the daily efficiency of bank teams.
ChatGPT for Financial Services includes high-quality financial data from providers like Daloopa, PitchBook, LSEG News, and Crunchbase, eliminating the need for manual MCP connector configuration. The data is uniformly indexed and hosted by OpenAI, enabling precise retrieval and granular citations that trace numbers and claims back to their original sources.
The platform natively supports advanced models, including GPT-6 Astra, and will automatically integrate new models as they are released. It features centralized management for access permissions and data connections, underpinned by the security and governance controls of ChatGPT Enterprise.
The initial collaboration focused on investment banking and equity research, where reliable data access and high-quality output generation are the most critical issues. OpenAI plans to use insights from these partnerships to refine the training and product, with future expansions into other financial service verticals.
In terms of data accessibility, OpenAI is working with S&P Capital IQ, LSEG, MSCI, Dow Jones Factiva, and Moody's to implement single sign-on (SSO) and authorization integration. This will allow users to access subscribed data seamlessly through their ChatGPT credentials.
The connector ecosystem already comprises over 50 integrations, including Datasite, Box, Preqin, FactSet, and Intapp. OpenAI has prioritized optimizing the most frequently used MCPs to ensure a reliable plug-and-play experience.
For security and compliance, the product is built on the foundation of ChatGPT Enterprise, utilizing SAML SSO, SCIM configuration, and role-based access control. It supports information barriers and multiple workspaces. Crucially, customer business data is never used to train OpenAI models, and all data transmission and storage are encrypted, with logging available for compliance audits.
An administrative portal allows for the deployment of standardized templates (Excel, Word, PowerPoint), ensuring that team members can quickly generate compliant valuation models, research notes, and pitchbooks.
The "ChatGPT for Financial Services" package is available to eligible institutions; interested parties should contact OpenAI's account team.