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)

Mobile: simplify leaderboard to Model + mean_22 (hide per-task/params/license columns, fixed layout so mean_22 stays visible)

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

Fix Pareto chart: model params as numeric millions (was string -> log10 NaN -> all points left)

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

Add IRT Ranking tab (theta ranking + per-task scores ordered by discrimination + 2-PL explainer); de-AI prose

Remove Raw Results link from header

Categories page: hover tooltips with category score + mean_16 + rank

Add /categories tab: 6 mini Pareto charts per task category (hash routing)

Footer links: add target=_blank to escape iframe; drop Results dataset

Sort task columns alphabetically by abbreviation

Submission URL: Space discussions instead of dataset

Add task list & sources link in footer

Fix maintainer name

Add Submit a model section + maintainer footer

HOTFIX TDZ: move shortLabel before labelPositions createMemo

Bigger chart 1100x620, all models labeled with collision-avoidance

Title -> MTEB Portuguese Leaderboard, version v1.0, mobile responsive

Move Pareto chart below table; soften Open/Closed badges

HOTFIX: TDZ error — move EXCLUDE/CLOSED_URLS before createMemo factories

Map closed-API model links to vendor docs (Gemini, OpenAI, Voyage, etc.)

Drop 2 incomplete models from table; add Open/Closed badge per row

Highlight only #1 per task (was: any score >= 0.7)

16 tasks (was misreporting 18)

Rewrite intro: drop unverifiable claim + over-technical detail

Centered header + intro paragraph; rename chart title to Open Pareto frontier

Add white halo to labels to mask Pareto line crossings

Compact chart: 30M-30B ticks, label only frontier+top5, drop incomplete models

Pareto chart: matplotlib-like style (white bg, labels, frontier dashed)

Add interactive Pareto chart, drop mean_18 toggle + tier badges

Initial leaderboard publish: 56 models × 18 tasks

initial commit