Card: refresh body to current state (93 models, 22 tasks/7 categories, mean_22, arXiv citation)

Card: list all 73 open-weight models so each shows the Space in 'Spaces using'

SEO: add card thumbnail + OG image/Twitter cards/canonical to the app

Rebrand headings to MTEB-BR (h1 'MTEB-BR Brazilian Portuguese Leaderboard'; title/desc/card MTEB(por) -> MTEB-BR)

Rebrand links MTEB-PT -> MTEB-BR (GitHub repo, HF org, discussions, issue templates); fix chart axis label

Upload README.md with huggingface_hub

cardData: point datasets to MTEB-BR/mteb-pt-results

Update leaderboard to MTEB(por) v2: 93 models x 22 tasks (mean_22)

16 tasks (was misreporting 18)

Initial leaderboard publish: 56 models × 18 tasks

initial commit