OpenAI's GPT-6.1 Terra Is Fast and Cheap but Trails Leading AI Models on Accounting Work, DualEntry Benchmark Finds

OpenAI has introduced GPT-6.1 Terra, an AI model whose accounting capabilities were assessed by the DualEntry Accounting AI Benchmark. The model achieved a score of 78.2%, ranking it 18th out of 22 leading AI systems evaluated.

The benchmark findings indicate that GPT-6.1 Terra performed strongest in specific accounting functions, namely transaction classification and journal entries. However, its results were weaker in other critical areas, including accounts payable and accounts receivable. This suggests the model's utility within accounting tasks may be segmented.

Compared to other prominent AI models, GPT-6.1 Terra's overall accounting proficiency trails. For example, Google's Gemini 4 Argon scored 91.1%, Zhipu AI's GLM-5.3 Flash and SpaceXAI's Grok 4.7 both achieved 88.1%, and OpenAI's own GPT-6.1 Sol performed better at 81.2%. These comparisons highlight that GPT-6.1 Terra is not currently a top performer for comprehensive accounting work as measured by this benchmark.

For businesses evaluating AI solutions, GPT-6.1 Terra's reported speed and cost, combined with its strengths in specific sub-tasks, might make it suitable for targeted automation. However, its weaker performance in areas like accounts payable and receivable suggests it may not serve as a standalone, end-to-end solution for broader or more complex financial processes, especially where consistent high accuracy across all accounting functions is required.