InVideo: Success Rate Tripled in Colorimetry with GPT-6 Astra

Key Takeaways

1The success rate for color calibration and correction has tripled with GPT-6 Astra

2Approximately 50 custom effects were created in one day by a few editors

3GPT-6 Astra enhances the planning and execution of complex edits while reducing the number of reasoning steps

Why it matters : These advancements allow professionals to save time on technical tasks while maintaining creative control.

InVideo attributes to GPT-6 Astra a threefold increase in success rates for color grading and correction, as well as the creation of around 50 effects in a single day. The company also describes improved planning for complex edits and the maintenance of editorial control by humans.

Tripled Success Rate and Accelerated Production

Sanket Shah, CEO of InVideo, claims that the success rate for color grading and correction tasks has tripled with GPT-6 Astra, whereas failure rates were previously very high. InVideo also reports the creation of approximately 50 custom effects in one day thanks to this model. The company emphasizes that, despite these advancements, professionals retain control over storytelling and creative choices while reducing the time spent on repetitive planning and frame-by-frame processing.

Colors: Choosing Approaches and Concrete Examples

According to InVideo, GPT-6 Astra stands out in color work that requires choosing between correction, grading, regeneration, using LUTs, or isolation. For example, modifying a background while preserving a person's complexion involves isolating and tracking the subject in the shots before applying color elsewhere. In this context, InVideo announces a threefold increase in success rates for these operations.

Le brief IA que lisent les pros

Le brief IA que les pros lisent chaque soir

Les 7 actus IA du jour, décryptées en 5 min. Gratuit.

Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.

Choisissez votre rythme

Gratuit · Pas de spam · Désabonnement en 1 clic

Planning and Executing Complex Edits

InVideo observes, with GPT-6 Astra, an improvement in the ability to plan and execute complex edits, as well as to place modifications in the right spots on the timeline. According to Sanket Shah, Astra requires fewer reasoning steps compared to previously evaluated models and utilizes fewer reasoning sequences and output tokens to accomplish complex tasks. The company also notes that the model can handle multiple instructions while maintaining the original objective, even as the task lengthens and increases in complexity. During this process, the agent must convert the editor's instructions into steps, choose the appropriate tools, perform the necessary operations, and control the outcome.

GPT-6 Astra Generates and Adapts Custom Effects for Each Sequence

InVideo explains that GPT-6 Astra allows for the transformation of descriptions and visual references into custom effects. The agent codes the effect to fit the sequences, inserts it onto the timeline, and creates controls for the editor to adjust it. This approach integrates into a video editor designed to handle heavy tasks while leaving the final decision to the human.