CFO says GPT-6 Sol hard-coded numbers in Excel, causing a major reconciliation error
Dawlphy · reddit · 2026-09-27
A CFO shared a real work mishap with GPT-6 Sol: while doing bank reconciliations, the model hard-coded a number into Excel instead of using =SUM, causing a significant calculation error, and added unnecessary sheets and labels that bloated the workbook. On the plus side, it flagged a discrepancy he hadn't prompted for. He runs a Codex folder with skills and deterministic operating rules for this exact task — GPT-5.6 Sol extra high currently reviews it — and says the new model could probably fill that role, but he doesn't understand its purpose or naming and likely won't use it for work again.
More from Models
- Fable 5.2 Leads MazeBench Token Efficiency; Billion-Token Runs Underway — patience_cave · 2026-11-16
- Fable 5.2 Tops MazeBench With 21% After 90 Hours and $6,000 in Credits — patience_cave · 2026-11-16
- Tencent Hy4 flash and InternLM Sage-S2.1-Flash both released, tool-calling tests requested — MajesticPuzzler · 2026-11-16
- User Shares Coral Reef Creation Labeled 'Qwen 4 Max (xhigh)', an Unannounced Model Name — intellectronica · 2026-11-16
- Stealth Model 'silicon-2' Surfaces in Devin, Suspected to Be Anthropic's Next Model — kimmonismus · 2026-11-16
- Google hides Gemini model label in AI Mode, rolls Gemini 4 Carbon to Pro subscribers — gaganghotra_ · 2026-11-16