More creative options and steadier sessions

You can now create across more image, voice, music, and text options while Ampere does a better job of preserving work through interruptions.

What's new

• Create and edit images with more choice. The image generator and editor now offer GPT Image 2.5 Flare and Sunburst. Choose a canvas size and quality for a new image, or edit source images with instructions and an optional mask.
• Make speech and reusable voices. Cartesia Sonic 3.6 can speak with a stock voice or one you have saved. With a recording you have permission to use, you can create an instant voice clone, listen to its audition, and reuse it from Custom Voices. MiniMax Music now uses Music 3.0 for songs with supplied or generated lyrics, or instrumental output.
• Choose a Gemini model for the task. Text generation and image analysis now offer Gemini 3.5 Flash-Lite, 3.1 Pro Preview, and 3.8 Flash. Flash-Lite remains the default; choose another option when your task calls for it.
• Continue more reliably after interruptions. Ampere can preserve a saved answer when a session is interrupted, let you respond to a recovered question, and accept a new follow-up after stopped work. Recovery depends on the evidence saved for that session.
• Plan content with a new guide. The Resource Hub now includes a worked guide to building a B2B SaaS or AI content strategy. Updated visuals across the guides make their examples easier to follow.

Why it matters

You have more ways to match the output to the job, whether you need a campaign image, a voiceover, music, or a text and image analysis model. More dependable session continuity also means less time reconstructing a brief when work is interrupted.

Where to find it

Open AI Tools to explore the updated image, speech, music, and Gemini tools. Visit Custom Voices to create and reuse a voice, start an Ampere session to work with the new choices, or read the content strategy guide in the Resource Hub.