13 New Image & Video Models
1. New generation models (6).png
🖼️ 6 New Image Models
We’ve gathered the image models drawing the most attention right now, all in one place. From realism to precise text to raw speed, pick the one that fits the job.
- Seedream 5.5 — ByteDance’s latest model, strong in lifelike images and stable multi-round editing. Even across dozens of successive edits it leaves untouched areas intact, and it handles small text and transparent backgrounds cleanly.
- Muse Image — Meta’s agentic image model that interprets the context of your prompt and plans the composition on its own, with particular strength in image editing.
- GPT Image 2.6 Flare — OpenAI’s speed-optimized model, built to turn out layout-driven work like posters and product imagery, fast.
- GPT Image 2.6 Sunburst — the detail-optimized variant, suited to dense, text-heavy compositions and work that needs consistency from one edit to the next.
- Qwen Image 3.0 — Alibaba’s image model, strong with long prompts and information-dense layouts, capturing fine detail like hair and texture along with small multilingual text.
- Qwen Image 3.0 Pro — the higher-performance variant of Qwen Image 3.0, producing more complex, more refined results.
🎞️ 7 New Video Models
Seven of the latest video models have been added to widen your range of direction. Combine them freely from reference-based generation to longer clips to quick drafts.
- Vidu — a reference-to-video model that turns anything — characters, scenes, props — into a reference for consistent video, with synced dialogue and sound effects.
- Kling 4.0 — the latest flagship, handling longer clips in a single generation. Keyframes, multi-reference inputs, and high-resolution output bring your intended direction to life with precision.
- Kling 4.0 Flash — the high-speed variant of Kling 4.0, handling text- and image-to-video and reference generation faster and lighter.
- Gemini Omni 1.2 Flash — Google’s latest Omni model that generates and edits video as if in conversation, with multi-turn editing and noticeably more accurate reference following.
- Wan 2.6 Flash — Alibaba’s speed-optimized, lightweight video model that handles image- and reference-based generation quickly — ideal for checking drafts.
- PixVerse V6 — strong in multi-shot generation that brings together camera work, character acting, and native audio.
- PixVerse V6 Max — the higher-end variant of PixVerse V6, delivering more polished video results.