Text (LLM), image, video, and multimodal—alongside the combined Model board.
Model leaderboard (all)
Performance inter-tâches (multimodal/vision/langage ; données d’exemple).
Cross-task overview; multimodal, vision, and language may split into additional columns or child boards once eval JSON is wired.
Public ranking policy: rows are sorted by composite score (desc). Composite score is a weighted sum of normalized sub-metrics; ties are broken by higher recent activity.
| Rang | Modèle | Éditeur / équipe | Type | Score | Notes |
|---|---|---|---|---|---|
| 1 | Demo-Vision-Pro | Demo Lab | Multimodal | 94.2 | Équilibre image + texte |
| 2 | NorthStar-MM | North AI | Multimodal | 92.8 | Fort en contexte long |
| 3 | Aurora-VL-7B | Aurora | Vision-langage | 91.5 | Adapté au edge |
| 4 | Helix-3 | Helix Research | Général | 90.1 | Appels d’outils stables |
| 5 | Kite-Small | Kite | Langage | 88.6 | Bon rapport qualité/prix |
| 6 | Lattice-R1 | Lattice | Raisonnement | 87.9 | Sous-scores maths/code élevés |
| 7 | Pulse-Audio-2 | Pulse | Multimodal parole | 86.4 | ASR/TTS combinés |
| 8 | Quark-Mini | Quark Systems | Langage | 85.2 | Faible latence |