Testing Gemini 4 Carbon on Creative Tasks

Bijan Bowen38m 36s
An illustration of testing a creative AI model.
0 comments · 0 votesOpen discussionClose discussion
Sign in to join the discussion

    The presenter explores Gemini 4 Carbon using a browser operating system, generated games, a 3D apartment and a robot-arm task. The demonstrations are informal and the comparisons reflect the prompts usedEvaluation is the systematic assessment of a system's behaviour or results against explicit criteria..

    An agent swarm is used to refine game outputA multi-agent system contains multiple AI agents that interact, coordinate, divide work, or influence one another while pursuing tasks., while the presenter describes a practical experiment generating a prototype car part and acknowledges a missed diagnosis.

    The presenter regards the visual results as an improvement over the preceding model, while noting incomplete outputs, performance issues and access and pricing constraints.

    Original YouTube thumbnailWatch on YouTube

    Video topic coverage

    • Agents & workflows52.7
    • Coding & tools84.6
    • Models & learning95.6
    • Infrastructure & compute0
    • Research & evaluation61.5
    • Safety & security0
    • Business & productivity0
    • Creative & media100
    • Robotics & perception26.4
    • Society & policy0

    Share this video

    An illustration of testing a creative AI model. With WWW.ARTIFICIAL-INTELLIGENCE.VIDEO, 10 November 2026 and duration 38m 36s.

    The presenter tests Gemini 4 Carbon on games, 3D scenes and a practical design task, with mixed results.