Canvas LLM Improvement

I would like to suggest upgrading the Large Language Model (LLM) used by Canvas AI.

Currently, the model appears to struggle with understanding technical documents and often provides incomplete or inaccurate answers to relatively straightforward questions. It also has limited ability to interpret visual content such as graphs, charts, and engineering plots, which are common in technical documentation.

An improved LLM with stronger reasoning and multimodal capabilities would significantly enhance the user experience by enabling it to:

• Provide more accurate and context-aware answers.
• Better understand engineering and technical terminology.
• Interpret graphs, charts, tables, and other visual elements within documents.
• Generate more comprehensive and reliable document summaries.
• Handle complex documents while maintaining context across multiple pages.

These improvements would make Canvas AI much more valuable for users working with engineering reports, operational documents, and technical studies, where understanding both text and visual information is essential.