Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.gmicloud.ai/llms.txt

Use this file to discover all available pages before exploring further.

On GMI Cloud, multimodal models are the media-focused APIs: they take and return non-text primary artifacts such as waveforms, video files, images, or 3D assets. They sit alongside LLM models, which are centered on language (and some vision-language understanding). Use this page as the section overview for those media modalities, then open the category that matches your output.

​
What is covered here

OutputRoleFull catalog
AudioTTS, voice cloning, musicAudio models
VideoT2V, I2V, editing, avatarsVideo models
ImageT2I, editing, batch jobsImage models

​
Technical topics (shared across media models)

  • Asynchronous jobs — Many video and image pipelines return a task you poll until completion; see tasks and artifacts in API References.
  • Inputs and controls — Prompts, reference images, first/last frames, masks, and duration or resolution caps are model-specific; each model page documents the supported schema.
  • Providers and tiers — Quality, latency, and cost vary by provider; marketplace and billing docs apply alongside per-model examples.
  • SDKs — Video and related flows may use the dedicated SDK, where documented under API References.

​
LLM vs multimodal

  • LLM pages focus on chat/completions-style usage (messages, sampling, optional tools or images in the request).
  • Multimodal (media) pages focus on file- or asset-centric flows: generate or edit media, download results, and handle longer-running jobs.
Some LLMs are vision-language models; they remain listed under LLM models because the primary integration pattern is still the language API.