OpenAI has launched ChatGPT for Financial Services, a tailored ChatGPT Work experience that combines built-in premium financial data with GPT-6 Astra's reasoning capabilities for investment banking and equity research teams. The product was shaped through a design partnership with Morgan Stanley and Evercore, targeting the data access and artifact creation pain points those firms identified as their biggest challenges.
The service includes datasets from Daloopa, PitchBook, LSEG News, and Crunchbase covering earnings transcripts, financial statements, company fundamentals, and private company data — all indexed and hosted on OpenAI infrastructure with granular citations that trace figures back to specific tables and passages. OpenAI says it will post-train models to interpret this data like experienced analysts. For firms with existing subscriptions, OpenAI is working with S&P Capital IQ, LSEG, MSCI, Dow Jones Factiva, and Moody's on shared sign-in and entitlement integrations so users can automatically access data they already pay for.
Confirmed
- GPT-6 Astra scores 69.9% on OfficeQA Pro (U.S. Treasury Bulletins with complex financial tables, charts, and footnotes), compared to 60.2% for GPT-5.6 Sol, according to OpenAI's internal testing.
- Built-in premium data from Daloopa, PitchBook, LSEG News, and Crunchbase available immediately with no separate contracts or connector setup.
- Optimized MCP connectors for S&P Global and FactSet, plus a broader ecosystem of 50+ connectors including Datasite, Box, Preqin, and Intapp.
- Admin publishing of Excel, Word, and PowerPoint templates for valuation models, research notes, and pitchbooks in firm format and style.
- Enterprise security: SAML SSO, SCIM provisioning, role-based access controls, data not used for training by default, encryption at rest and in transit, configurable workspace retention, compliance log exports via OpenAI Compliance Platform, and multiple workspaces for information barriers.
- Available to eligible financial institutions; interested firms should contact OpenAI or their account team.
Unknown
- Pricing for ChatGPT for Financial Services — not disclosed in the announcement.
- Independent replication of the OfficeQA Pro benchmark (69.9% for GPT-6 Astra vs 60.2% for GPT-5.6 Sol) — vendor-reported only.
- Rollout timeline beyond "available to eligible financial institutions" — no general availability date or phased rollout schedule provided.
- Scope of post-training on financial datasets — cadence, coverage expansion roadmap, and measurable impact on hallucination rates for financial figures not specified.
- Entitlement integration completeness with S&P Capital IQ, LSEG, MSCI, Dow Jones Factiva, and Moody's — described as "working on" rather than live.
- Whether GPT-6 Astra's higher OfficeQA Pro score translates to measurable time savings or error reduction in real banking workflows.
Our take
OpenAI is betting that vertical integration — hosting premium data natively, optimizing connectors, and post-training models on financial corpora — creates a defensible moat against generic ChatGPT Enterprise plus third-party plugins. The design partnership with Morgan Stanley and Evercore signals top-tier banks are willing to co-develop rather than just buy, but the lack of pricing and independent benchmarks means the value proposition remains unproven for the broader market. The real test will be whether firms see enough workflow consolidation to justify a dedicated SKU over their existing ChatGPT Enterprise + MCP stack.
