Add MoveBench Release 1 leaderboard

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+domain_scores.png filter=lfs diff=lfs merge=lfs -text

README.md

@@ -1,10 +1,24 @@
 ---
 title: MoveBench Leaderboard
-emoji: 🚀
-colorFrom: gray
-colorTo: blue
+emoji: 🏃
+colorFrom: blue
+colorTo: green
 sdk: static
 pinned: false
+license: cc-by-4.0
+short_description: Sports and health coaching benchmark leaderboard
+models:
+  - zai-org/GLM-5.3
+  - moonshotai/Kimi-K3
+  - MiniMaxAI/MiniMax-M3
+  - tencent/Hy3
+datasets:
+  - KeepaceAI/MoveBench
+tags:
+  - leaderboard
+  - benchmark
+  - sports
+  - health
 ---
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+Leaderboard for [MoveBench](https://huggingface.co/datasets/KeepaceAI/MoveBench) Release 1 (80 public questions).

index.html

@@ -1,19 +1,50 @@
 <!doctype html>
-<html>
-	<head>
-		<meta charset="utf-8" />
-		<meta name="viewport" content="width=device-width" />
-		<title>My static Space</title>
-		<link rel="stylesheet" href="style.css" />
-	</head>
-	<body>
-		<div class="card">
-			<h1>Welcome to your static Space!</h1>
-			<p>You can modify this app directly by editing <i>index.html</i> in the Files and versions tab.</p>
-			<p>
-				Also don't forget to check the
-				<a href="https://huggingface.co/docs/hub/spaces" target="_blank">Spaces documentation</a>.
-			</p>
-		</div>
-	</body>
-</html>
+<html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
+<title>MoveBench Leaderboard</title>
+<style>
+:root{--bg:#fcfcfb;--fg:#0b0b0b;--muted:#5f5e5a;--line:#e4e3de;--accent:#2a78d6}
+@media (prefers-color-scheme:dark){:root{--bg:#1a1a19;--fg:#f0efec;--muted:#c3c2b7;--line:#3a3a37;--accent:#3987e5}}
+body{margin:0;background:var(--bg);color:var(--fg);font:15px/1.55 system-ui,-apple-system,"PingFang SC",sans-serif}
+main{max-width:960px;margin:0 auto;padding:32px 16px 48px}
+h1{font-size:28px;margin:0 0 4px} .sub{color:var(--muted);margin:0 0 16px}
+nav a{color:var(--accent);margin-right:14px;text-decoration:none}
+.wrap{overflow-x:auto;margin:20px 0} table{border-collapse:collapse;width:100%;font-variant-numeric:tabular-nums}
+th,td{padding:8px 10px;border-bottom:1px solid var(--line);text-align:right;white-space:nowrap}
+th:nth-child(2),td:nth-child(2){text-align:left} th{color:var(--muted);font-weight:600;font-size:13px}
+td a{color:var(--accent);text-decoration:none}
+img{max-width:100%;border-radius:8px;border:1px solid var(--line)}
+ul{padding-left:20px} li{margin:4px 0} .note{color:var(--muted);font-size:13px}
+</style></head><body><main>
+<h1>MoveBench Leaderboard</h1>
+<p class="sub">Chinese sports and health coaching benchmark · Release 1 · 80 public questions</p>
+<nav><a href="https://huggingface.co/datasets/KeepaceAI/MoveBench">Dataset</a><a href="https://arxiv.org/abs/2610.26185">Paper</a><a href="https://github.com/Keep-Tech/MoveBench">GitHub</a><a href="https://github.com/Keep-Tech/MoveBench/blob/main/docs/SUBMISSION.md">Submit a model</a></nav>
+<div class="wrap"><table>
+<thead><tr><th>Rank</th><th>Model</th><th>Overall</th><th>OQ (50)</th><th>MCQ (30)</th><th>PG</th><th>DA</th><th>QA</th></tr></thead>
+<tbody>
+<tr><td>1</td><td>Claude Opus 5.5</td><td><b>0.897</b></td><td>0.892</td><td>30/30</td><td>0.860</td><td>0.927</td><td>0.947</td></tr>
+<tr><td>2</td><td>Qwen3.8-Max</td><td><b>0.886</b></td><td>0.880</td><td>30/30</td><td>0.839</td><td>0.932</td><td>0.938</td></tr>
+<tr><td>3</td><td><a href="https://huggingface.co/zai-org/GLM-5.3">GLM-5.3</a></td><td><b>0.878</b></td><td>0.872</td><td>30/30</td><td>0.825</td><td>0.925</td><td>0.947</td></tr>
+<tr><td>4</td><td><a href="https://huggingface.co/MiniMaxAI/MiniMax-M3">MiniMax-M3</a></td><td><b>0.864</b></td><td>0.859</td><td>29/30</td><td>0.803</td><td>0.911</td><td>0.949</td></tr>
+<tr><td>5</td><td><a href="https://huggingface.co/moonshotai/Kimi-K3">Kimi-K3</a></td><td><b>0.860</b></td><td>0.853</td><td>30/30</td><td>0.804</td><td>0.904</td><td>0.936</td></tr>
+<tr><td>6</td><td><a href="https://huggingface.co/tencent/Hy3">HY3</a></td><td><b>0.858</b></td><td>0.851</td><td>30/30</td><td>0.811</td><td>0.896</td><td>0.919</td></tr>
+<tr><td>7</td><td>Doubao-Seed-2.1-Pro</td><td><b>0.857</b></td><td>0.850</td><td>30/30</td><td>0.803</td><td>0.904</td><td>0.925</td></tr>
+<tr><td>8</td><td>GPT-6</td><td><b>0.848</b></td><td>0.842</td><td>29/30</td><td>0.802</td><td>0.897</td><td>0.897</td></tr>
+<tr><td>9</td><td>Gemini-4-Argon</td><td><b>0.838</b></td><td>0.830</td><td>30/30</td><td>0.776</td><td>0.894</td><td>0.913</td></tr>
+<tr><td>10</td><td>MiMo-V2.6-Pro</td><td><b>0.837</b></td><td>0.830</td><td>30/30</td><td>0.778</td><td>0.878</td><td>0.924</td></tr>
+</tbody></table></div>
+<h2>Key findings</h2>
+<ul>
+<li><b>Plan generation is the hardest domain for every model.</b> PG scores range from 0.776 to 0.860, compared with 0.878–0.932 on DA and 0.897–0.949 on QA.</li>
+<li><b>The top 10 models are within 6 points.</b> In a bootstrap over the 50 open-ended questions, Claude Opus 5.5 stays in 1st–2nd place (90% interval), but the order from 4th place down is not stable.</li>
+<li><b>Multiple choice is nearly saturated.</b> 8 of 10 models answer all 30 correctly.</li>
+</ul>
+<img src="domain_scores.png" alt="Domain scores for the 10 leaderboard models; every model scores lowest on plan generation">
+<h2>How scores are computed</h2>
+<ul class="note">
+<li>Scores range from 0 to 1: points earned divided by points possible.</li>
+<li>Open-ended questions are scored against expert rubrics by two LLM judges, and the two scores are averaged. Multiple-choice questions are scored by exact match.</li>
+<li>Scores come from each model's full 809-question run, keeping only the 80 public questions. They are not the paper's full-benchmark scores.</li>
+<li>Rubrics and reference answers have been public on GitHub, so results may be affected by data contamination. Treat them as indicative.</li>
+<li>To be listed, submit your model's answers following the submission guide. Maintainers score and review submissions before publishing.</li>
+</ul>
+</main></body></html>