Overview
- Adobe introduced Firefly Image Model 5 with native 4‑megapixel output, improved human rendering and layered, prompt‑based editing, which is available in beta.
- Firefly adds Prompt to Edit for conversational image adjustments, with layered editing demonstrated as a scene‑aware system that treats objects as editable elements.
- New audio and video features arrive with Generate Soundtrack and Generate Speech in public beta, ElevenLabs voices included, and a web‑based Firefly video editor in private beta with a waitlist.
- Creators will be able to train custom Firefly models on their own assets, with closed beta underway, a November waitlist and broader rollout planned toward year‑end.
- Adobe is rolling out AI assistants across apps—Express available now and Photoshop in private beta—expanding partner model options such as Google Gemini 2.5 Flash and Black Forest Labs FLUX.1 Kontext, and developing ChatGPT control for Express with OpenAI.