gen-ai-media

name	gen-ai-media
description	AI model selection guidance for media generation — which image, video, TTS, and music model to use for each use case. Triggers on "which model", "best model for", "image generation model", "video model", "TTS provider", or cost optimization questions.

AI Media Model Selection Guide

Image Generation Models

Model	Via	Cost	Best For	Quality
Ideogram V3 Turbo	Replicate	$0.02/img	General-purpose, stylized images	High, stylized
GPT Image 1.5 Low	Replicate	$0.013/img	Product cutouts, simple items	Good, photorealistic
GPT Image 1.5 Medium	Replicate	$0.05/img	Compositions, character designs	Very good
GPT Image 1.5 High	Replicate	$0.14/img	Hero images, avatar frames	Best, full fidelity
FLUX.1 Kontext Pro	Replicate	$0.04/img	Budget avatar frames with reference	Good, fast
FLUX.1 Fill Pro	Replicate	$0.05/img	Outpainting / background extension	Seamless extension
Nano Banana (Gemini 2.5 Flash)	Google AI	$0.15/img	Alternative composition	Good
Bria BG Removal	Replicate	$0.001/img	Background removal	Clean edges

Decision Tree

• Need stylized/artistic? → Ideogram V3 Turbo ($0.02)
• Need photorealistic product? → GPT Image 1.5 (Low for cutouts, Medium for compositions)
• Need character with expressions? → GPT Image 1.5 High ($0.14) or FLUX Kontext ($0.04 budget)
• Need to extend background? → FLUX.1 Fill Pro ($0.05)
• Need transparent cutout? → Generate + Bria BG Removal ($0.001)

Video Generation Models

Model	Via	Cost	Best For
Seedance 1.5 Pro 480p	Replicate	$0.013/sec (no audio), $0.025/sec (with audio)	Cost-efficient reels
Seedance 1.5 Pro 720p	Replicate	$0.026/sec (no audio), $0.052/sec (with audio)	Higher quality
Seedance 1 Lite 480p	Replicate	$0.018/sec	Legacy, avoid
prunaai/p-video	Replicate	~$0.07/shot	Audio-driven animation (talking characters)
PixVerse Lipsync	Replicate	$0.08325/sec	Post-production lip-sync

Decision Tree

• Image-to-video (standard)? → Seedance Pro 480p without audio ($0.013/sec)
• Need native lip movements? → Seedance Pro with audio ($0.025/sec)
• Need audio-driven character animation? → p-video (~$0.07/shot)
• Need to add lip-sync to existing video? → PixVerse Lipsync ($0.083/sec)

TTS Models

Provider	Model	Cost	Best For
MiniMax	Speech 2.8 HD	~$0.000005/char	Budget voiceovers, marketing content
ElevenLabs	Turbo v2.5 / Multilingual v2	~$0.30/1000 chars	Premium ads, character voices

Decision Tree

• Marketing reels/content? → MiniMax (99% cheaper)
• Paid ads requiring premium quality? → ElevenLabs
• Character voices with high expressiveness? → ElevenLabs (low stability, high style)

Music Models

Service	Cost	Best For
Lyria 2 (Google)	$0.0001/sec	AI-generated custom music
Jamendo	Free (CC)	Pre-made royalty-free tracks

Background Removal (Video)

Model	Cost	Output
Robust Video Matting	$0.046/run (flat)	Alpha mask → ffmpeg merge
Bria Video	$0.14/sec	Native transparent WebM

• Default to RVM (30x cheaper). Use Bria only for hero content.