fix(xai): route video models by modality

This commit is contained in:
Julien Talbot
2026-06-01 19:00:30 -07:00
committed by Teknium
parent eee32cdd52
commit 8104b20269
8 changed files with 167 additions and 19 deletions
+2 -1
View File
@@ -182,7 +182,8 @@ The `x_search` toolset auto-enables whenever xAI credentials (a SuperGrok / X Pr
| Chat | `grok-4.20-multi-agent-0309` | Multi-agent variant |
| Image | `grok-imagine-image` | Default; ~510 s |
| Image | `grok-imagine-image-quality` | Higher fidelity; ~1020 s |
| Video | `grok-imagine-video` | Text-to-video and image-to-video; up to 7 reference images |
| Video | `grok-imagine-video` | Text-to-video |
| Video | `grok-imagine-video-1.5-preview` | Image-to-video; dated alias `grok-imagine-video-1.5-2026-05-30` |
| TTS | (default voice) | xAI `/v1/tts` endpoint |
The chat catalog is derived live from the on-disk `models.dev` cache; new xAI releases appear automatically once that cache refreshes. `grok-4.3` is always pinned to the top of the list.