ChatGPT vs Claude
ChatGPT from OpenAI and Claude from Anthropic are AI assistants that can help finance teams work with text, documents and spreadsheets. Both accept instructions in ordinary language and can support drafting or analysis of supplied material. A useful comparison tests how each product handles the team's actual work, from the first instruction through to a result that a colleague can check.
Useful tests include finding a clause in a prospectus, reconciling a table or explaining a change to an internal policy. The choice also depends on which data the firm may share and the terms of its agreement with the provider. Testing both products on approved material shows how much checking and correction each task requires before a result can enter a customer or compliance workflow.
The two products and their makers
ChatGPT is made by OpenAI. ChatGPT runs in a browser, on the desktop and on the phone, takes files, searches the web, writes and executes code in a sandbox, and connects to other software. The model line from OpenAI carries the GPT name, and the company sells a separate coding product, OpenAI Codex.
Claude is made by Anthropic and runs in the same places. The Claude model line carries the names Opus, Sonnet and Haiku, and the company sells Claude Code for developers. Both companies also sell API access, which is how either model ends up inside software a bank already owns.
Testing documents, spreadsheets and drafts
Both products accept files, and ChatGPT works with documents and spreadsheets. Claude can also create and edit files, including spreadsheets and presentations. Available features depend on the account and the settings approved by the firm.
Test long documents with questions about specific clauses and ask for source locations. Then check the original document to see whether the cited passage supports the answer. For spreadsheets, require traceable formulas and calculation steps, and compare results with a known reference value.
For drafting, assess factual accuracy and whether the language suits its audience. Multi-step workflows need additional approvals and logs, as with AI agents in finance.
Codex and Claude Code in development
OpenAI Codex and Claude Code help developers work with a repository. They can read code, propose changes and run commands. Codex approvals and Claude Code permissions are configurable; neither product universally asks for confirmation before every file edit.
A firm should define permitted directories, network access and approval rules. Test both on the same development task, inspect the diff and run the existing test suite. Access to the runtime environment and traceable changes matter alongside the code produced.
Data rules and contract terms
Before use, establish which data classes may be processed, whether inputs may be used for training and where processing takes place. Record which data is excluded.
Procurement and legal teams should review liability, logging, audit rights and termination terms. Duties under the EU AI Act also depend on the use case. Approval for internal drafting does not automatically permit use in credit decisions.
How does a Frankfurt team test both before deciding?
A team can test both products with its own documents and a scoring sheet written before the test. Take ten real tasks the team does every month, run them through both products, and score each answer on whether it is correct, whether it cites a source you can check, and how much rework it needed. This gives the team evidence from its own workflows.
Do it on an account whose terms allow the data you are using, and keep the outputs. When procurement asks why one product was chosen, that record is the answer. AI tools for finance lists what else belongs on the shortlist, since the two products here are not the only options.
Model choice and Finance Loop
Finance Loop brings together people who have compared ChatGPT and Claude at a bank to discuss the factors behind their decisions. Finance Loop hosts this exchange in Frankfurt without taking a vendor's side.
Finance Loop is a professional network supporting the institutional use of emerging technologies in finance, such as AI, tokenization, stablecoins, and DeFi. Finance Loop helps its members build skills and personal networks in these fields: Investment & Digital Assets, Payments & Digital Money, Digital Infrastructure & Sovereignty, and Risk & Compliance.