AI
Hebbia finds OpenAI's GPT-6 Astra boosts financial research
In Hebbia's tests, OpenAI's GPT-6 Astra led the best alternative by 13.6 percentage points on citing required source types, and 99.5% of its citations were relevant.

Key Takeaways
- 1Hebbia found GPT-6 Astra scored 13.6 percentage points above the best alternative tested on citing the required types of sources.
- 2GPT-6 Astra met 97.9% of presentation requirements, leading competitors by 14.4 percentage points.
- 399.5% of the excerpts GPT-6 Astra cited were relevant to the question.
- 4In Hebbia's workflow, GPT-6 Astra selects supporting documents and turns the analysis into native PowerPoint slides.
Hebbia found that OpenAI's GPT-6 Astra significantly improves financial research, scoring 13.6 percentage points above the best alternative tested on citing the required types of sources and meeting 97.9% of presentation requirements in tests conducted between August and September 2026. This performance helps streamline complex financial workflows for professionals.
The findings, reported by OpenAI, highlight Astra's capabilities in accurately finding required sources and generating high-quality presentations. Financial institutions rely on precise data and professional outputs for critical decisions.
Hebbia’s Matrix platform aims to integrate these AI advancements, linking answers directly to their source documents. This enhances traceability and user confidence in AI-generated insights.
What are the benefits of GPT-6 Astra for financial research?
GPT-6 Astra offers significant benefits for financial research by excelling in source selection and presentation quality. In Hebbia's evaluations, it cited the required types of sources more often and adhered better to presentation guidelines compared to other models, which helps financial professionals create accurate and polished documents faster.The model achieved a 13.6 percentage-point lead over the best alternative in citing required source types. Notably, 99.5% of the excerpts cited by GPT-6 Astra were relevant to the questions posed. This precision is critical when reviewing thousands of documents, such as for a strategic acquisition.
Hebbia also noted Astra's superior performance in presentation generation. The model met 97.9% of applicable presentation requirements, leading the next best alternative by 14.4 percentage points.
We’ve been strong partners with OpenAI over the course of Hebbia’s lifetime, and the leaps GPT-6 Astra has made in terms of quality outputs and accuracy have truly been surprising. GPT-6 Astra shines compared to any other model we’ve tested before on these two fronts.
— Kavita Anand, Member of Technical Staff, Product, Hebbia
Streamlining Complex Financial Workflows
GPT-6 Astra streamlines complex financial workflows by automating document analysis and presentation creation, allowing professionals to focus on strategic assessment rather than manual tasks. This includes reviewing large data rooms and converting research into formatted client deliverables.In Hebbia’s hypothetical example, a banker assessing an acquisition with roughly 2,000 files could load the data room into Matrix to examine customer concentration, contract terms, and churn, then turn that research into a diligence summary and valuation deck in the firm’s template.
Hebbia wants bankers to spend less time reformatting decks and checking sources, and more time assessing a deal. Learn more about advanced models like GPT-5.4 Mini and Nano on AI Gateway.
| Metric | GPT-6 Astra Performance | Lead Over Best Alternative |
|---|---|---|
| Required Source Types Cited | Not stated | 13.6 percentage points higher |
| Citation Relevance | 99.5% of cited excerpts relevant | Not stated |
| Presentation Requirements Met | 97.9% applicable requirements met | 14.4 percentage points higher |



