Kumo Auto
Default route for task diagnosis, context-aware routing and cost-aware output.
Auto routeCost-aware
Kumo Lumo
Choose by task, not by memorizing model names.
Kumo Lumo organizes model routes around real work: lightweight questions, long context, research, writing, image generation, voice scripts and video prompt planning.
GPT-5.4
Efficient reasoning, writing, summaries, code explanation and everyday structured work.
ReasoningWriting
Claude-Opus-4.8
Long-form reasoning, document review, careful writing and high-context synthesis.
Long contextReview
Gemini-3.0
Multimodal analysis, image-aware tasks and cross-file context work.
MultimodalContext
GPT Image-2
Image prompts, generation planning, edits and visual production specifications.
ImageVisual spec
Eleven Lab
Voiceover scripts, narration pacing, pronunciation notes and audio-ready output.
VoiceNarration
| Route | Best for | Point profile | Input types |
|---|---|---|---|
| Kumo Auto | Unknown tasks, routing, budget-aware answers | Auto | Text, files, images, context |
| GPT-5.4-Mini | Lightweight chat, simple summaries, quick drafts | Low | Text, context |
| Claude-Opus-4.8 | Long writing, careful analysis, document work | High | Text, files, context |
| Gemini-3.1-Pro | Long-context and multimodal synthesis | High | Text, files, image context |
| Nano-Banana-Pro | Creative image directions and rapid visual variants | Low to medium | Text, image prompts |
| Seedance, Kling, Veo, Sora, Runway | Video prompt planning and generation briefs | Medium to high | Text, image prompts, video briefs |