Models
The models behind Sume's video, image, speech, and music endpoints: ids, clip lengths, resolutions, aspect ratios, audio, references, and migration.
Start with What AI models does Sume support? Every model id, by API surface
FLUX.2 native-language prompts: Japanese or Korean vs English
BFL says native-language FLUX.2 prompts often look more culturally authentic. Send both versions to a Sume FLUX.2 id with n and compare the results.
Veo 3.1 Lite preview shuts down Oct 22: Omni Flash on Sume
Google lists veo-3.1-lite-generate-preview shutting down October 22, 2026 with gemini-omni-1.1-flash as replacement. Sume id: gemini-omni-flash-1.1.
Grok Imagine Video 1.5 text-to-video at 1080p: xAI vs Sume
xAI lists native 1080p text-to-video for grok-imagine-video-1.5. Sume's row is image-to-video only at 480p and 720p. How the two differ and what to send.
HappyHorse 1.1 API ids: 480P to 1080P, 3 to 15 seconds
Alibaba lists happyhorse-1.1-t2v, -i2v and -r2v at 480P, 720P or 1080P for 3 to 15 seconds. Sume's catalog docs name HappyHorse 1.1 as not in the v1 catalog.
Ideogram 4 custom_model_uri vs Sume catalog ids and sume/auto
Ideogram's custom-model endpoint picks a private model by custom_model_uri. Sume has no such field: you choose a catalog id or sume/auto. What to send instead.
OpenAI organization verification for GPT Image 2.5 and Sume ids
OpenAI says you may need API Organization Verification before using GPT Image models. Sume lists the same two models as openai/gpt-image-2.5 and -sunburst.
Recraft V4 Styles: 10 references, and the Sume Recraft id
Recraft V4 Styles takes one to 10 reference images in Precise or Flexible mode. Sume lists recraft/recraft-v4 as text-to-image only and rejects references.
Suno v6 change one lyric: what Sume Music does instead
Suno v6 can swap one lyric or line and keep the rest. Sume Music is brief-driven and documents no lyric editing, so you edit the brief and generate a new track.
Sync react-1 model_mode and emotion vs Sume TTS emotion and speed
Sync react-1 edits an existing video with model_mode lips, face or head plus an emotion prompt. On Sume, emotion and speed are set in the TTS audio request.
wan2.2-animate-move vs animate-mix, and Sume's motion transfer
animate-move drives a still character with a reference video; animate-mix swaps a character into a video. Sume's Genjutsu row covers the move-style case.
wan2.7-videoedit camera replication vs Sume's video_url edit
Alibaba names wan2.7-videoedit for effect and camera-movement replication. On Sume, video edit is the video_url field on the Gemini Omni Flash 1.1 id.
YouTube fhd, qhd, uhd thumbnails in the API: a 3840 AI image
YouTube's API notes fhd, qhd and uhd thumbnail keys for some videos. Sume's ChatGPT Image 2.5 image_size accepts a 3840 edge, enough for a 4K image.
Cartesia Sonic 2 and Turbo stop October 20: Sume model ids
Cartesia retires sonic-2 and sonic-turbo on October 20, 2026. Sume's TTS Router lists sonic-3.6, 3.5, 3, latest and preview; here is what to send instead.
ElevenLabs STT edit instructions vs Sume transcribe: what Sume returns
ElevenLabs added transcript-editing instructions up to 2,000 characters to its speech to text. Sume's transcribe endpoint returns a plain transcript.
ElevenLabs dialogue continuity vs Sume TTS segments: what differs
ElevenLabs Text to Dialogue links takes with previous_text and request ids. Sume TTS has no such fields; it returns gapless sentence segments instead.
Lyria 3 Clip vs Lyria 3.5: which one runs on Sume?
Google lists Lyria 3 Clip as a 30-second model and Lyria 3.5 as minutes long. Sume's Music Router routes lyria-3.5, lyria-3-pro or sume/music-auto.
Lyria 3.5 takes 131,072 tokens; Sume music prompts, 5,000 chars
Google lists a 131,072 token input limit for Lyria 3.5. Sume's music prompt field takes 1 to 5,000 characters. What fits and how to trim.
Lyria 3.5 takes up to 10 images; Sume music takes one image_url
Google's Lyria guide allows up to 10 images as input. Sume's music docs describe one optional public HTTPS image_url. What that changes.
Can I distribute AI music through TuneCore? Suno vs Sume tracks
Suno says tracks from its partner model can go out through Believe and TuneCore. That is a Suno feature. What it means if your tracks come from Sume.
Veo 3.1 only makes 16:9 and 9:16; which Sume video models add more
Google's Veo 3.1 supports 16:9 and 9:16. On Sume, Seedance, MiniMax and Wan add 4:3, 1:1 and 3:4, Kling adds 1:1, and Grok takes no ratio.
Veo 3.1 prompts cap at 1,024 tokens; Sume's Omni at 20,000 characters
Google caps a Veo 3.1 text prompt at 1,024 tokens. On Sume, gemini-omni-flash-1.1 documents a 20,000-character cap. Tokens and characters differ.
Which Sume video models accept an audio reference?
Seven Sume video models take reference audio: the four Seedance 2.x ids, Wan 3.0 and the two MiniMax H3 models. Kling, Grok and Gemini Omni Flash do not.
Which Sume video models accept a reference video?
Nine Sume video ids take a video input: the four Seedance 2.x ids, Wan 3.0, both MiniMax H3 models, Omni Flash and Genjutsu. Kling and Grok do not.
Anam Cara-4 anime and 3D avatar images vs Sume's photo input
Anam says Cara-4 animates anime and 3D character images. Sume's avatar API takes a prompt, profile or public HTTPS reference image and promises no style.
Anam Cara-4 Director Notes vs Sume scene prompts and silence
Anam Director Notes steer a live avatar with a style and an Expressivity control. Sume directs scenes with a prompt or photo, silence beats and a quality tier.
Anam Cara-4 portrait 768x1152 vs Sume 9:16 avatar at 720p
Anam Cara-4 renders natively at 768x1152 portrait. Sume's avatar video defaults to 9:16 at a fixed 720p, with plans estimated at 4-60 seconds.
AssemblyAI Universal-3.6 Pro 32 languages vs Sume STT language_code
AssemblyAI streaming defaults to Universal-3.6 Pro with 32 languages; 3.5 Pro has 19. Sume STT 1.0 takes one optional language_code hint and auto-detects.
Sonic 3.6 written disfluencies: pacing and Sume's emotion guide
Sonic 3.6 shifts pacing for written hesitations like uh. How to write a natural-sounding transcript, and what Sume's optional emotion guide adds.
Cartesia STT keyterms for brand names, and Sume STT without keyterms
Cartesia STT adds keyterm prompting for brand names and invented words. Sume STT has no keyterm field, so here is how to fix brand names after transcription.
Deepgram Nova-3 de-CH, Lithuanian, Portuguese: language_code on Sume
Deepgram improved Nova-3 for Swiss German, Lithuanian and Portuguese on September 24. What to send as language_code on Sume STT for regional variants.
FLUX 3 Action is a robot-control model, not an image model
FLUX 3 Action is an open-weight 7B world action model for robots. It predicts actions, not images, so it is no drop-in for the FLUX ids Sume serves.
FLUX 3 Image bounding boxes vs Sume masked edit: what each takes
FLUX 3 Image takes a JSON list of boxes at the end of the prompt. Sume's image edit takes a mask image: mask_image_url with image_urls. Differences, no overlap.
FLUX 3 Image 10 reference images vs Sume's 1-10 image_urls
FLUX 3 Image takes up to ten references named by prompt position. Sume's image edit takes one to ten public image_urls, with no FLUX 3 id in the docs.
FLUX 3 x mimic: one backbone for video, audio and robot actions
BFL says FLUX 3 is one model trained jointly on images, video and audio, and FLUX-mimic adds robot actions. What that does and does not tell a media API user.
FLUX 3 video continuation caps at 15 s: extending a clip on Sume
BFL cut FLUX 3 video continuation (v2v) to 15 seconds on Aug 17; other modes stay at 20. Sume has no continuation mode: chain a last frame instead.
FLUX 3 Video 4K uhd (3840x2176) and which Sume video models do 4K
FLUX 3 Video's uhd resolution is 3840 x 2176 at 16:9. On Sume, a 4K request goes to a model whose supported_resolutions lists it; here is what the docs say.
Gemini 3.8 Live Extended Thinking: the id, and Sume's audio ids
gemini-3.8-live-extended-thinking is a Live API audio-to-audio model id. Sume has no live audio route; its audio ids are sume/music-auto and TTS Router ids.
Grok Imagine 21:9 and 5:2 ratios vs Sume's Grok aspect list
xAI added 21:9 and 5:2 to Grok Imagine image on Aug 28. Sume's Grok list has 20:9 and 2:1 but neither of those two, so pick a listed ratio for banners.
Genjutsu Restyle's 20+ styles vs Sume motion transfer inputs
Higgsfield Genjutsu Restyle picks from 20+ styles in its app. Sume's motion transfer takes a source video plus 1 to 8 reference images, with no style list.
Kling 4.0 on-screen text vs Sume burned-in captions
Kling 4.0 renders on-screen text inside the video. Sume burns captions onto a finished clip; language is a speech-to-text hint, and cues add authored text.
Kling 4.0 video extension, forward or backward, done by hand on Sume
Kling 4.0 extends a video forward or backward in one workflow. On Sume you chain clips yourself: pull a frame, pass it as first_frame or last_frame.
Kling Avatar output length follows audio; Sume H3 Max is 5-14.8 s
On fal, Kling Avatar v2 Pro output duration matches the audio file. On Sume the audio sets length too, within route windows from 5 to 14.8 s up to 60 s.
Kling Avatar v2 Standard vs Pro: fal tiers vs Sume quality
fal prices Kling Avatar v2 Pro at $0.115/s and Standard at $0.0562/s. Sume Avatar Video has its own quality values: standard, plus and max.
MiniMax H3 Max reference audio needs an image or video on Sume
Krea says MiniMax H3 Max audio references cannot be sent alone; Sume's reference_audio_urls has the same rule. Limits for images and videos compared.
Wan 3.0 references or start/end frames, not both: two Sume calls
Krea says a Wan 3.0 job takes reference media or start/end frames, not both. Sume lists them as separate modes, so split the work into two requests.
LTX-2.5 Cinemagraph LoRA vs image-to-video on Sume
The LTX-2.5 Cinemagraph LoRA is an image-to-video adapter for selective motion. On Sume, start from first_frame and describe the motion in the prompt.
Luma layers API: 10 RGBA layers vs Sume's flat image result
Luma type layering splits one image into up to 10 ordered RGBA PNG layers. Sume returns flat images at data[].url; n counts images, not layers.
Luma Scenes Ray 3.2 or Seedance 2: pin a model or use auto
Luma says Ray 3.2 and Seedance 2 are not interchangeable. On Sume, pin a catalog model for a known clip length, or send sume/auto and let Sume pick.