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
Which BPM makes a 3 to 10 second Omni clip a whole number of bars?
Gemini Omni Flash 1.1 clips run 3 to 10 seconds. At 120 BPM a 10 s clip is 5 bars; at 128 BPM no whole-second length fits. A table of the BPMs that line up.
Does the Sume spend cap cover the agent's own model tokens?
No: Sume's cap excludes the agent's LLM turn; debited_usd_micros includes it. At list rates that turn is about $0.004 on Haiku 5.5 and $0.40 on Astra.
A 19-second single shot: which Sume video models take it in one job
Wan 3.0 and Seedance 2.5 take 19 s in one job ($4.75 and $27.01 at 1080p). Seedance 2.0, H3 and Omni top out at 15 s, 15 s and 10 s, so they need two jobs.
8:1 strips and 21:9 banners: which image models take them
Nano Banana 2.1 lists 8:1, 4:1, 1:4 and 1:8 on Sume; Pro stops at 21:9 and 9:16. Prices per image for each ratio choice, read 2026-10-08.
Agent Completion or Format run: where can you pick Haiku 5.5?
Only a Format run takes a catalog model id. An Agent Completion accepts model sume-agent and nothing else, so Haiku 5.5 is a Format-run choice.
Can a Haiku 5.5 agent look at images? Sume's 30-image attachment rules
Anthropic says all current Claude models take image input. Sume accepts up to 30 images, 30 MB each, 500 MB total; Haiku 5.5 is a Format-run pick.
Which Sume video models accept a music track as an audio reference?
Seedance 2.x, Wan 3.0, MiniMax H3 and H3 Max take audio references; Gemini Omni Flash 1.1 does not. Prices for 8 s plus a $0.125 bed.
ChatGPT Image 2 vs 2.5 at medium quality, 16:9: price on Sume
At 16:9 and medium quality, ChatGPT Image 2 bills $0.0525 per image on Sume and ChatGPT Image 2.5 bills $0.0105: five times less. Table, batch math and a swap.
Image-to-video prompt: write the motion, not the photo
When a photo is already pinned as the first frame, spend the prompt on motion, camera and sound, not on describing what the picture shows. A worked rewrite.
Nano Banana 2 is retired on Sume: same price on Nano Banana 2.1?
Sume retired google/nano-banana-2 and runs old requests as Nano Banana 2.1 at the same price card: $0.10 at 1K billed. Tiers, ids and what the job stores.
Replace one prop in a video with AI: bottle to apple edit prompt
Swap one object in a finished clip with gemini-omni-flash-1.1 on Sume: the docs example prompt, fields you cannot set, and an 8-second price.
MiniMax H3 Max on Sume: video, lip-sync and recast, which id to call
Three Sume ids carry the MiniMax H3 Max name: minimax-h3-max for video, a lip-sync route and h3-max-recast. What each takes, its window and its rate.
Video edit prompt: say what stays, then what changes (Omni Flash)
A prompt pattern for Gemini Omni Flash 1.1 video edit on Sume: one change per pass, an explicit keep clause, and what the video_url route fixes for you.
Which Sume image models are text-to-image only, with no photo edits
Five Sume image models reject reference images: Higgsfield Soul, Imagen 4 Fast, Imagen 4 Ultra, Recraft V4 and Qwen Image Max. Prices, ratios and edit options.
Do low, medium and high mean the same on every Sume image model?
No. On Sume only ChatGPT Image 2, 2.5 and Ideogram 4.5 accept low, medium and high, and the tiers change price in different ways. Defaults and the tier ladders.
Fix extra fingers in an AI image: a masked GPT Image 2.5 edit
Fix a bad hand in an AI image with a mask_url edit on openai/gpt-image-2.5: mask only the hand, keep the rest, and budget about $0.0835 per try at high.
OpenAI names Flare or Sunburst to replace gpt-image-1: which to pin?
OpenAI's deprecations page lists gpt-image-2.5-sunburst or -flare for gpt-image-1. Sume serves both at one price, so the choice is quality. A script to A/B.
Remove an object from a photo without a mask: Ideogram 4.5 prompt edit
Remove a bin, cable or stray person from a photo with no mask: send it to ideogram/ideogram-v4.5 and name the object. $0.0375 to $0.275 per image on Sume.
Square 1:1 ad video: which Sume models accept it, and 4:5
Seedance, Kling, Wan and MiniMax accept 1:1 on Sume; Gemini Omni takes only 16:9 and 9:16. No video model lists 4:5, so Feed needs a crop. Table and steps.
Gemini Omni video edit on Sume: video_url price by output second
Edit a clip with gemini-omni-flash-1.1 by sending video_url and a prompt. Sume bills $0.125 per second at 720p default, so a 6-second edit is $0.75.
Minimum clip length by Sume video model: 2, 3, 4 or 5 seconds
Wan 3.0 starts at 2 s, Gemini Omni Flash 1.1 at 3 s, Seedance and Genjutsu at 4 s, MiniMax H3 and Recast at 5 s. Full min and max table for a Sora port.
Seedance pixel sizes by ratio on Sume: 864x496 at 480p, 1920x1080
Sume prices Seedance by video tokens counted from fixed pixel sizes per ratio and resolution. The size table, and why 480p 16:9 is 864x496 and not 854x480.
Silent clips after Sora: which Sume models take generate_audio false
Gemini Omni Flash 1.1 rejects generate_audio false; Kling 3 prices audio on at $0.21 a second against $0.14 off; recast and motion transfer keep source sound.
Wan 3.0 document-to-video claim vs what Sume accepts as reference
Alibaba says Wan 3.0 reads PDFs, slides and docs. Sume's catalog lists image, video and audio references for it, not documents. Read the field before you plan.
Two reference images, one clip: Omni's cat-and-yarn pattern on Sume
Google's docs show two images, a cat and yarn, producing one video. Here is the Sume request with IMAGE_REF tokens and what changes with a brand product.
MAI-Transcribe-2-Streaming 2.5% WER: what the test audio mix is
The 2.5% word error rate comes from a chunked-streaming index: 50% AA-AgentTalk, 25% VoxPopuli, 25% Earnings22. Here is how that maps to a recorded call.
A 60-second Seedance 2.5 piece: two 30 s jobs and one last frame
Together lists multi-round extension for Seedance 2.5; Sume documents none. Build 60 s as two 30 s jobs and a last-frame handoff: $34.68 at 720p, $7.50 on Wan.
Music Router image_url: public HTTPS only, and null to clear it
Sume music image_url must be a public HTTPS image, and null clears it on a reused request object. When a mood still helps and when the prompt does more.
Native 4K or 2K image edits: pixel ceilings by model
How many pixels each model reaches: Flux 3 Image near 16.8 MP, GPT Image 2.5 capped at 8,294,400 on Sume, Ideogram 4.5 at 2K. What Sume serves.
Music Router prompt limit: 5,000 characters, and what to put in them
Sume Music Router prompts run 1 to 5,000 characters, with no duration field or negative prompt; length goes in the text. A brief budget and a checker.
MiniMax H3 or H3 Max after Sora: native 768p vs latent 1080p prices
On Sume, H3 renders native 480p or 768p and bills 2K/4K upscales; H3 Max adds 1080p as a latent refinement of 768p. 10-second prices side by side.
Three characters, three dances: Omni IMAGE_REF and VIDEO_REF tokens
Google's Omni 1.1 demo swaps three dancers for a dog, an octopus and a bear. Here is the same request on Sume, with the 0-based reference tokens in order.
Does Wan 3.0 do 4K? Claim check against vendor pages and Sume
Alibaba lists 480p, 720p and 1080p for Wan 3.0, Morphic says no 4K, and Sume rejects 4k for wan-3.0. Where 4K exists on Sume and at what clip length.
gemini-3.1-flash-image is deprecated: move to Nano Banana 2.1
Google deprecated gemini-3.1-flash-image on Oct 6 with no shutdown date and named gemini-nano-banana-2.1 as the replacement. What to change, and what Sume does.
Nano Banana 2.1 has no 512px size in Google's guide; Sume lists 512
Google's guide says 512px is unsupported on Nano Banana 2.1, yet Sume's catalog enum for it starts at 512. How to test what a call returns.
Nano Banana 2.1 references: 10 objects, 4 characters vs Pro 6, 5, 3
Google splits Nano Banana 2.1 references into 10 objects and 4 characters, and Pro into 6 objects, 5 characters and 3 style images. How Sume's flat list maps.
Is Gemini 3.8 Flash TTS in the Live API? No, per Google's model page
Google's model page marks the Live API unsupported for Gemini 3.8 Flash TTS. For live voice use Gemini 3.8 Live; Sume TTS is for async audio jobs.
Where can you use Gemini 3.8 Flash TTS today? API, Notebook, Vids
Google lists Gemini 3.8 TTS in the Gemini API, AI Studio, Gemini Notebook and Google Vids now; the Enterprise API is coming soon. What Sume offers instead.
HappyHorse 1.1 on Cloudflare: 2,500-character prompt limit
Cloudflare's HappyHorse 1.1 text-to-video page caps the prompt at 2,500 characters and clips at 3 to 15 seconds. HappyHorse is not in the Sume catalog.
MAI-Voice-2.1 has three style lists: emotion, role and expressive
MAI-Voice-2.1 styles come in three vocabularies: 19 emotions, 6 roles and 11 expressive tags. Which voices use the expressive set, and how Sume differs.
MAI-Voice-2.1 whispering and shouting: the 19 voices that list both
Only 19 MAI-Voice-2.1 voices list whispering and shouting by our count. English-UK and Korean voices do not. Sume has no whisper switch.
MAI-Voice-2.1 has 97 voices: en-US has 7, es-ES and en-AU have 1 each
Microsoft's MAI-Voice-2.1 doc lists 97 prebuilt voices over 28 locales. We counted per locale: some have seven choices, others only one.
Gemini voice remixing is 'coming soon': what Sume TTS can change today
Google says Gemini voice remixing for timbre, pitch, pace and accent is coming soon. Sume TTS today has speed, volume and an emotion string only.
MiniMax H3 outputs 24 fps and 32 kHz stereo: what Sume's H3 adds
MiniMax's H3 open-source post lists 24 fps video and 32 kHz stereo audio, 4 to 15 s clips and 768p by default. Sume's minimax-h3 takes 5 to 15 s.
Qwen Image 2.1 sizes (2752x1536, 2400x1792) vs Sume Qwen ratios
The Qwen-Image-2.1 card lists 2048x2048, 2400x1792 and 2752x1536. Sume qwen-image takes 13 aspect_ratio values instead of pixels. How they line up.
Which Sume video models take an audio_url reference, which do not
Sume video rows list supported_input_references per model. The docs show seedance-2 with audio_url, H3 Max with audio, and Gemini Omni Flash 1.1 without it.
MAI-Voice-2.1 neutral-only voices: Hungarian, Romanian, en-IN
Several MAI-Voice-2.1 voices list only a neutral style: five Hungarian, four Romanian and both en-IN voices. Check the style list before you cast an ad.
Can Gemini Omni Flash speak my script? It takes no audio input on Sume
On Sume, Gemini Omni Flash 1.1 takes image and video references but no audio reference. It makes its own sound. For your script in a face, use avatar video.