How the term is used here In this channel's coverage of GPT-6.1 Soul, illustrative tasks refers to the work the model does when it produces visual output that helps explain or communicate an idea — maps, dashboards, and presentation decks. The reviewer highlights this as one of the model's clear strengths: its visual design output is described as a step up over previous models, creating material that actually illustrates ideas rather than just filling space. Why it matters in the comparison During the live vibe check comparing GPT-6.1 Soul against Astra, Soul, and Opus, illustrative tasks are treated as evidence of the model's improved visual understanding — standing in contrast to its honest weak spot, natural writing, where output reads flat and segmented. The takeaway: pick this model for design and visual communication work, not for prose. - Illustrative tasks = producing maps, dashboards, and decks that visually explain ideas. - GPT-6.1 Soul is described as notably strong at these tasks. - This strength is contrasted with the model's weakness in writing.