Docs · Routing
Picking the right model
A strengths cheat sheet across every video, image, voice, and music model on Varosity. Use it as a static reference, or query the same logic via POST /v1/route.
Let your agent pick automatically
POST /v1/route takes a shot description and returns a ranked recommendation filtered to your configured providers.
79 models
veo-3.1
Veo 3.1 Fast
cinematic · WaveSpeed
- +best lip sync
- +synced audio
- +talking heads
$0.15/s · Varosity Credits only
veo-3.1-quality
Veo 3.1 (Quality)
cinematic · WaveSpeed
- +state-of-the-art fidelity
- +best lip sync
- +native audio
$0.4/s · Varosity Credits only
veo-3.1-lite
Veo 3.1 Lite
cinematic · WaveSpeed
- +cheapest Veo tier
- +native audio
- +high-volume iteration
$0.05/s · Varosity Credits only
kling-3.0
Kling 3.0 Pro
cinematic · fal.ai
- +multi-shot consistency
- +cinematic motion
- +dialogue close-ups
$0.168/s BYOK · $0.168/s Credits
flux-3-max
FLUX 3 Max
audio-native · fal.ai
- +native audio
- +5–20s clips
- +promotional price until Nov 7
$0.0425/s BYOK · $0.0425/s Credits
vidu-q4
Vidu Q4
audio-native · fal.ai
- +native audio
- +3–16s clips
- +low price
- −needs a reference image (image-to-video only)
$0.0665/s BYOK · $0.0665/s Credits
grok-imagine-video
Grok Imagine Video
audio-native · fal.ai
- +cheapest frontier video with native audio
- +synchronized sound + dialogue
- +flexible 1–15s duration
$0.07/s BYOK · $0.07/s Credits
seedance-1.0-pro
Seedance 1.0 Pro
cinematic · fal.ai
- +multi-shot from one prompt
- +1080p by default
- +single takes up to 12s
- −no audio — use seedance-2.0, veo-3.1 or kling-3.0 for sound
$0.1215/s BYOK · $0.1215/s Credits
seedance-2.0
Seedance 2.0
cinematic · fal.ai
- +top-ranked motion + physics
- +cinematic camera control
- +native synced audio
$0.3024/s BYOK · $0.3024/s Credits
seedance-2.0-fast
Seedance 2.0 Fast
cinematic · fal.ai
- +near-flagship quality at lower cost
- +faster renders
- +native synced audio
$0.2419/s BYOK · $0.2419/s Credits
omnihuman
OmniHuman 1.5
avatar · fal.ai
- +talking avatars
- +from single photo
- +lip-sync from audio
- −needs photo + audio
$0.16/s BYOK · $0.16/s Credits
fal-live-portrait
LivePortrait (Performance Transfer)
avatar · fal.ai
- +performance transfer from a driving video
- +transfers head motion + facial expressions + lip movement
- +onto a different person's photo
- −face & head only (not full body/hands)
$0.05/s BYOK · $0.05/s Credits
fal-wan-animate
Wan 2.2 Animate (Full-Body Transfer)
creator · fal.ai
- +FULL-BODY motion transfer from a driving video
- +transfers whole-body movement + face + expression
- +onto a different character image
- −~3-4 min render
$0.16/s BYOK · $0.16/s Credits
kling-3.0-replicate
Kling 3.0 Pro (Replicate)
cinematic · Replicate
- +multi-shot consistency
- +cinematic motion
- +alternate route when fal is busy
- −about twice the price of kling-3.0 for the same model
$0.336/s BYOK · $0.336/s Credits
happyhorse-1.0
HappyHorse 1.0
audio-native · fal.ai
- +#1-ranked motion + prompt adherence
- +joint audio-video generation
- +multilingual lip-sync
$0.14/s BYOK · $0.14/s Credits
luma-ray-3-fal
Luma Ray 3.2 (fal)
cinematic · fal.ai
- +high-fidelity motion
- +strong realism
- −540p output at this price
$0.1/s BYOK · $0.1/s Credits
runway-gen-4.5-replicate
Runway Gen-4.5 (Replicate)
creator · Replicate
- +camera control
- +scene consistency
- +prompt adherence
- −5s or 10s clips only
$0.12/s BYOK · $0.12/s Credits
elevenlabs-tts
ElevenLabs Multilingual v2
tts · ElevenLabs
- +29 languages
- +emotion control
- +voice cloning
- −streaming latency vs Cartesia
$0.0012/s BYOK · $0.0012/s Credits
elevenlabs-tts-v3
ElevenLabs v3 (Alpha)
tts · ElevenLabs
- +dialogue style
- +highest expressiveness
- −alpha — quality varies
$0.0012/s BYOK · $0.0012/s Credits
elevenlabs-sts
ElevenLabs Speech-to-Speech
tts · ElevenLabs
- +voice conversion
- +preserves delivery/performance
$0.002/s BYOK · $0.002/s Credits
elevenlabs-dubbing
ElevenLabs Dubbing
tts · ElevenLabs
- +translate + dub to 30+ languages
- +speaker-aware
$0.0084/s BYOK · $0.0084/s Credits
elevenlabs-scribe
ElevenLabs Scribe (Speech-to-Text)
tts · ElevenLabs
- +transcription
- +speaker diarization
- +word-level timestamps
$0.0000611/s BYOK · $0.0001/s Credits
elevenlabs-music
ElevenLabs Music
music · ElevenLabs
- +commercially licensed training data
- +low-latency
- +fast
- −fewer genres than Suno
$0.0025/s BYOK · $0.0025/s Credits
lyria-2
Google Lyria 2
music · fal.ai
- +instrumental detail
- +short 30 s cues
- −no vocals
$0.10/song (any length)
flux-1-schnell
FLUX.1 [schnell]
creator · fal.ai
- +fast (about 5 s)
- +good prompt adherence
- +low cost
- −no negative prompts
$0.003/image BYOK · 1¢/image Credits
flux-3-klein
FLUX.3 Klein
creator · fal.ai
- +Black Forest Labs' newest image model
- +fast and cheap
- +built-in prompt expansion
- −text-to-image only via Varosity
$0.009/image BYOK · 1¢/image Credits
krea-3
Krea 3
creator · fal.ai
- +photoreal and stylized looks
- +built-in prompt expansion
- +about 1 megapixel
- −text-to-image only via Varosity
$0.025/image BYOK · 3¢/image Credits
flux-2-pro
FLUX.2 [pro]
creator · fal.ai
- +current-gen FLUX quality
- +strong prompt adherence
- −pricier than schnell
$0.03/image BYOK · 3¢/image Credits
flux-2-klein
FLUX.2 [klein]
creator · fal.ai
- +current-gen FLUX at draft price
- +crisper text than FLUX.1
- +fast (4-step)
- −below flux-2-pro fidelity
$0.006/image BYOK · 1¢/image Credits
z-image-turbo
Z-Image Turbo
creator · fal.ai
- +ultra-cheap (~$0.01)
- +~1s render
- +open-weight
- −draft quality — not for hero shots
$0.01/image BYOK · 1¢/image Credits
seedream-4
Seedream 4.0
creator · fal.ai
- +photorealism
- +strong composition
- +low cost
$0.03/image BYOK · 3¢/image Credits
flux-1.1-pro
FLUX 1.1 Pro
cinematic · Replicate
- +highest-quality Flux
- +strong prompt adherence
- +fine detail
- −slower than schnell
$0.04/image BYOK · 4¢/image Credits
grok-imagine-image
Grok Imagine Image
creator · fal.ai
- +cheap ($0.02/image)
- +photorealism
- +fast
$0.02/image BYOK · 2¢/image Credits
gemini-3-pro-image
Nano Banana Pro (Gemini 3 Pro Image)
creator · Google AI Studio
- +best-in-class legible in-image text
- +multilingual text
- +reasoning-driven composition
- −pricier than the other Google image tiers
$0.134/image BYOK · 14¢/image Credits
gemini-3.1-flash-image
Nano Banana 2 (Gemini 3.1 Flash Image)
creator · Google AI Studio
- +production-scale quality at flash speed
- +legible in-image text
- −video-to-image (its unique trick) not wired via Varosity yet
$0.067/image BYOK · 7¢/image Credits
nano-banana-2.1
Nano Banana 2.1
creator · WaveSpeed
- +newest Nano Banana: design, subject consistency, prompt adherence
- +legible in-image text
- +edits from a reference image
- −1K via Varosity
4¢/image · Varosity Credits only
seedream-5.0
Seedream 5.0 Pro
creator · WaveSpeed
- +newest Seedream: strong prompt following
- +edits from a reference image
- +wide range of shapes
- −1K via Varosity
5¢/image · Varosity Credits only
seedream-5.0-flash
Seedream 5.0 Flash
creator · WaveSpeed
- +fast, cheaper Seedream 5.0
- +2K output
- +edits from a reference image
- −less detail than Seedream 5.0 Pro on busy scenes
3¢/image · Varosity Credits only
qwen-image-3.0
Qwen Image 3.0
creator · WaveSpeed
- +2K images for 3¢
- +strong prompt understanding
- +edits from a reference image
3¢/image · Varosity Credits only
grok-imagine-image-2.0
Grok Imagine Image 2.0
creator · WaveSpeed
- +xAI's newest image model
- +2K photoreal images
- +edits from a reference image
8¢/image · Varosity Credits only
gpt-image-2.5-flare
GPT Image 2.5 Flare
creator · WaveSpeed
- +OpenAI's newest image model
- +five quality tiers from 1¢
- +edits from a reference image
- −1K output only through Varosity
3¢/image · Varosity Credits only
gpt-image-2.5-sunburst
GPT Image 2.5 Sunburst
creator · WaveSpeed
- +OpenAI's newest image model, precision tier
- +extra fidelity on intricate detail
- +edits from a reference image
- −slower than Flare
3¢/image · Varosity Credits only
gemini-3.1-flash-lite-image
Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image)
creator · Google AI Studio
- +ultra-low latency
- +cheapest Google tier
- +high-volume iteration
- −lower detail ceiling than Nano Banana 2 / Pro
$0.034/image BYOK · 4¢/image Credits
ideogram-v3
Ideogram V3
creator · fal.ai
- +best text-in-image
- +legible typography
- +poster / cover art
- −aspect ratios limited to fixed presets
$0.06/image BYOK · 6¢/image Credits
recraft-v3
Recraft V3
creator · fal.ai
- +long detailed prompts
- +brand-consistent style
- +vector-style output
- −slower than Flux Schnell (~10–20s)
$0.04/image BYOK · 4¢/image Credits
gpt-image-2
GPT Image 2
creator · OpenAI
- +best-in-class instruction following
- +accurate in-image text
- +photorealistic detail
- −rate-limited per OpenAI plan
$0.08/image BYOK · 8¢/image Credits
nano-banana
Nano Banana
creator · Replicate
- +very fast
- +good identity preservation
- +image editing
- −less photoreal than Flux Pro
$0.04/image BYOK · 4¢/image Credits
suno-v4
Suno V6 (via muapi)
music · muapi.ai
- +best vocal generation
- +genre + mood tags
- +full song structure
- −review licensing before commercial use
$0.10/song · Varosity Credits only
muapi-flux-dev
Flux Dev
creator · muapi.ai
- +12B parameter model
- +strong prompt adherence
- +fast guided distillation
- −slower than Flux Schnell
2¢/image · Varosity Credits only
muapi-wan-effects
WAN Video Effects
creator · muapi.ai
- +named effect catalog (Cakeify, Squish, VHS, Samurai…)
- +frame consistency
- +platform-funded
- −short clips only (≤10s)
$0.06/s · Varosity Credits only
muapi-latentsync
LatentSync Lip-Sync
avatar · muapi.ai
- +smooth temporal consistency
- +fast inference
- +any video + audio
- −needs pre-existing video
$0.05/s · Varosity Credits only
muapi-wan-t2v
WAN 2.1 Text-to-Video
creator · muapi.ai
- +platform-funded (no BYOK)
- +up to 720p / high quality
- +reliable fallback
- −5-second clips only
$0.06/s · Varosity Credits only
openai-tts-1
OpenAI TTS-1
tts · OpenAI
- +fast
- +9 voices
- +low latency
- −slightly lower quality than TTS-1 HD
$0.000225/s BYOK · $0.0002/s Credits
openai-tts-1-hd
OpenAI TTS-1 HD
tts · OpenAI
- +highest OpenAI voice quality
- +9 voices
- +natural prosody
- −2× cost of TTS-1
$0.00045/s BYOK · $0.0005/s Credits
deepgram-aura-2
Deepgram Aura 2
tts · Deepgram
- +ultra-low latency
- +natural prosody
- +cheap (about $0.027 a minute)
- −fewer voices outside English (es, de, fr, nl, it, ja)
$0.00045/s BYOK · $0.0005/s Credits
cartesia-sonic-3.6
Cartesia Sonic 3.6
tts · Cartesia
- +~90ms TTFA
- +42 languages
- +AI laughter + emotion
- −voice ids are library-specific — call list_voices
$0.000975/s BYOK · $0.001/s Credits
cartesia-sonic-3
Cartesia Sonic 3
tts · Cartesia
- +~90ms TTFA
- +42 languages
- +AI laughter + emotion
- −voice ids are library-specific — call list_voices
$0.000975/s BYOK · $0.001/s Credits
fal-kokoro
Kokoro (fal)
tts · fal.ai
- +open-weight (MIT)
- +runs on CPU
- +very cheap
- −fewer expressive controls than ElevenLabs
$0.0003/s BYOK · $0.0003/s Credits
fal-chatterbox
Chatterbox (fal)
tts · fal.ai
- +open-weight (MIT)
- +expressive + emotion control
- +instant voice cloning
- −voiceId is a clone-sample URL, not a library id
$0.0006/s BYOK · $0.0006/s Credits
fish-audio-tts
Fish Audio
tts · Fish Audio
- +multilingual
- +cheap
- +large community voice library
- −voice ids are library-specific — call list_voices
$0.000225/s BYOK · $0.0002/s Credits
heygen-avatar-4
HeyGen Avatar (Digital Twin)
avatar · HeyGen
- +studio-grade lip sync
- +Avatar IV / Avatar V motion engines
- +voice emotion + speed control
- −requires a pre-trained avatar look id
$0.08/s BYOK · $0.08/s Credits
heygen-photo-avatar
HeyGen Photo Avatar (Avatar IV)
avatar · HeyGen
- +animate ANY photo as the speaker
- +Avatar IV motion engine
- +motion prompt + expressiveness control
- −needs a clear front-facing photo
$0.08/s BYOK · $0.08/s Credits
heygen-cinematic
HeyGen Cinematic Avatar
cinematic · HeyGen
- +prompt-driven cinematic shots
- +blends 1–3 avatar looks into a scene
- +reference videos/images for style
- −4–15s per clip
$0.1/s BYOK · $0.1/s Credits
heygen-video-agent
HeyGen Video Agent
avatar · HeyGen
- +prompt → finished video
- +agent writes script, picks avatar & scenes
- +accepts reference files
- −least granular control
$0.12/s BYOK · $0.12/s Credits
heygen-video-translate
HeyGen Video Translate
avatar · HeyGen
- +multilingual lip-sync dubbing
- +preserves original speaker appearance
- +supports 40+ languages
- −requires source video with clear speech
$0.1/s BYOK · $0.1/s Credits
d-id-talks
D-ID AI Presenter
avatar · D-ID
- +talking avatars from any photo
- +text-to-presenter
- +fast render
- −fixed framing
$0.05/s BYOK · $0.05/s Credits
hunyuan-video
Hunyuan Video
cinematic · fal.ai
- +open-source quality
- +long coherent motion
- +strong physics
- −fixed ~5s clips
$0.09/s BYOK · $0.09/s Credits
ltx-video
LTX Video
creator · fal.ai
- +fastest open video model (<5s)
- +image-to-video
- +good for iteration
- −lower detail than Kling/Veo
$0.04/s BYOK · $0.04/s Credits
wan-2.6
Wan 2.6
creator · fal.ai
- +open-weight (Apache-2.0)
- +strong motion
- +self-hostable
$0.1/s BYOK · $0.1/s Credits
hunyuan-video-1.5
HunyuanVideo 1.5
cinematic · fal.ai
- +open-weight (Apache-2.0)
- +long coherent motion
- +strong physics
- −slower render
$0.09/s BYOK · $0.09/s Credits
ws-luma-ray-2
Luma Ray 2
cinematic · WaveSpeed
- +fluid motion
- +cinematic quality
- +strong prompt adherence
- −higher cost than Pika
$0.08/s · Varosity Credits only
ws-pika-2.2
Pika 2.2
creator · WaveSpeed
- +fast generation
- +stylized output
- +good character consistency
- −shorter max duration
$0.04/s · Varosity Credits only
ws-hailuo-02
Hailuo 02
cinematic · WaveSpeed
- +physics realism
- +complex motion
- +high resolution
- −fixed 6s duration
$0.0833/s · Varosity Credits only
ws-kling-4.0
Kling 4.0
cinematic · WaveSpeed
- +newest Kling, native audio
- +single takes up to 30s
- +21:9 widescreen
- −slow: a 5s clip can take about 10 minutes
$0.157/s · Varosity Credits only
wan-3.0
Wan 3.0
cinematic · WaveSpeed
- +newest Wan, with audio
- +single takes of 2 to 30 seconds
- +text- or image-to-video
- −720p via Varosity
$0.1/s · Varosity Credits only
minimax-h3
MiniMax H3
cinematic · WaveSpeed
- +MiniMax's newest video model, with sound
- +4 to 15 second takes
- +21:9 widescreen
- −768p via Varosity
$0.1/s · Varosity Credits only
grok-imagine-video-1.5
Grok Imagine Video 1.5
cinematic · WaveSpeed
- +newest Grok Imagine, with sound
- +1 to 15 second takes
- +stylized looks
- −text-to-video only via Varosity
$0.14/s · Varosity Credits only
seedance-2.0-mini
Seedance 2.0 Mini
cinematic · WaveSpeed
- +Seedance 2.0 at half the Fast price
- +animates a start frame, with sound
- +4 to 15 second takes
- −image-to-video only (needs a start frame)
$0.12/s · Varosity Credits only
ws-runway-gen4
Runway Gen 4
cinematic · WaveSpeed
- +camera control
- +cinematic motion
- +image-to-video
- −needs a reference image (image-to-video only)
$0.056/s · Varosity Credits only
Smart Route API
Skip the cheat sheet. Pass a shot description to /v1/route and get a ranked recommendation filtered to your configured providers.
// Request
POST https://varosity.ai/api/v1/route
Authorization: Bearer vsk_…
Content-Type: application/json
{
"shot_description": "close-up of a barista pouring milk, dialogue, cinematic",
"modality": "video",
"duration_s": 8,
"budget_cents": 200
}// Response
{
"ok": true,
"primary": {
"model_id": "veo-3.1",
"vendor": "fal",
"billing_mode": "byok",
"estimated_cost_cents": 120,
"reasoning": "Dialogue + close-up + cinematic → lip-sync wins",
"confidence": 0.95
},
"fallbacks": [
{ "model_id": "seedance-1.0-pro", "billing_mode": "byok",
"reasoning": "Similar motion quality; no native audio" }
],
"filtered_out": [
{ "model_id": "kling-3.0-replicate", "reason": "no_byok_key" }
]
}