Current AI media models — September 2026

Google flagship video generation with native audio; Standard tier.

Faster Veo 3.1 tier with native audio.

Lowest-cost Veo 3.1 tier for 720p/1080p.

Fast multimodal video generation/editing with native audio.

Current Kling generation model with multi-shot storytelling and Elements consistency.

Kling multimodal model for advanced references, composition and edits.

Fast Kling 3.0 generation route.

Latest Seedance route in major developer catalogs, with image/video/reference inputs.

Reference-driven cinematic Seedance generation.

Faster Seedance 2.0 route.

Lower-cost Seedance tier.

Current Wan 3.0 general video API.

Faster Wan 3.0 Prime; multimodal, up to 30s.

Current HappyHorse family for cinematic motion and cross-clip consistency.

MiniMax H3 multimodal video generation with 768p/2K output.

Fast H3 Max route available in Runway Dev.

Runway flagship generation model with professional/HDR output options.

Fast image-to-video Runway model.

Prompt-guided video-to-video editing model.

Performance/character motion transfer workflow.

xAI video model accepting text, image and audio inputs.

Current Luma video model with frame-level control and HDR/EXR workflows.

Fast LTX 2.3 generation route.

Higher-quality LTX 2.3 with audio/video-conditioned workflows.

Current Pika general text/image-to-video model.

Released Sep 18, 2026; cinematic generation with automatic scene direction and references.

Released Sep 17, 2026; transfers motion from source video.

New Sep 2026 HiDream model; current public API is image-conditioned with dynamic or fixed 10s output.

Adobe native video generation model inside Firefly.

Professional video generation model available in current creator suites.

Current Vidu Q3 general flagship with synchronized audio and smart scene cuts.

Fast/current Vidu Q3 route with audio, scene cuts and reference-to-video.

Reference-to-video Q3 model focused on balance, consistency and synchronized audio.

PixVerse flagship general model with audio, multi-clip generation and reference workflows.

Film-production PixVerse model for action choreography, storyboard/reference consistency and transitions.

Tencent current HY-Video-1.5; official replacement for the retiring legacy Hunyuan video APIs.

Released Sep 15, 2026; real-time interactive avatar/video model with voice interaction and reference-image control.

Released Sep 15, 2026; real-time video-stream editing with style/clothing/character/background references.

Current fast Google image generation/editing model.

Lower-cost Gemini 3.1 Flash Lite image tier.

Higher-quality Google image model still active in current creative suites.

High-quality GPT Image 2.5 route.

Faster GPT Image 2.5 route.

Current default Midjourney image version (since Jul 24, 2026).

Current Ideogram frontier image model.

Current Recraft raster image model.

Runway image generation model.

Fast Runway image model.

Low-cost Runway image generation route.

Current xAI image generation route.

Current high-quality Seedream route in developer catalogs.

Lower-cost Seedream 5 route.

Current Luma image/multimodal model; 2K output.

Higher-quality Luma UNI tier; 2K output.

Current Adobe native image model in Photoshop/Firefly.

Current FLUX.2 Pro model in major creative suites.

Released Sep 2026 for campaign image generation/editing.

Higgsfield image generation model.

Current recommended Qwen image generation/editing model for complex layouts and typography.

Current balanced Qwen 3.0 image generation/editing model.

Current high-quality Wan image route; up to 4K and consistency-focused workflows.

Current lower-cost Wan 2.7 image route.

Tencent current HY-Image-3.0; official replacement for legacy Hunyuan image APIs.

Google flagship full-song music generation model.

Audio generation route in Runway Dev.

Current ElevenLabs speech generation route via Runway Dev.

Sound-effect generation.

Pika text-to-music model.

Pika sound-effect model.

Current MiniMax music generation route.

Kling text/video-to-audio route.

Released Sep 11, 2026; current most advanced Eleven Music model with composition plans and reference audio.

Released Sep 9, 2026; Suno current flagship music model.

Experimental v6 variant for more varied/unpredictable music generations.

Fast/light v6 variant available across Suno plans.