AI リーダーボード
総合および各カテゴリのトップモデル。YouTubeやXで使われているグラフの元データとなる公開ベンチマークです。独自スコアの公開は行いません：すべての数値は産出元のベンチマークに帰属し、取得日とともに出典を明示しています。
Real people are shown the same prompt answered by two anonymous models side by side and vote for the better answer. Millions of these blind head-to-head votes are fed into a Bradley-Terry statistical model (the successor to the Elo system it started with) which converts win/loss pairs into a single rating per model. A higher rating means people picked that model more often against strong opposition. This snapshot uses the 'style control' variant, which is the site's default: it statistically adjusts for answer length and formatting so a model cannot climb simply by writing longer, prettier replies.
制限事項: It measures which answer people LIKE, not which answer is CORRECT - a confident, well-written wrong answer can still win a vote. Voters are self-selected volunteers rather than a representative sample, prompts skew toward what that crowd chooses to type, and models with few votes have wide confidence intervals (ci_low/ci_high) that often overlap the models ranked above and below them. Treat small rank gaps as ties.
出所: LMArena (formerly LMSYS Chatbot Arena) ↗ · 公開済み 2026-08-11 · 基礎データ ↗ · Dataset released under Creative Commons Attribution 4.0 (CC BY 4.0). Reuse is permitted with attribution - credit LMArena and link to the leaderboard.
| # | モデル | 組織 | Arena score (human preference) |
|---|---|---|---|
| 1 | claude-fable-5 | Anthropic | 1554 95% CI 1545 to 1562 5,751 votes |
| 2 | claude-opus-4-7-thinking | Anthropic | 1552 95% CI 1546 to 1558 17,231 votes |
| 3 | claude-opus-4-6-thinking | Anthropic | 1552 95% CI 1546 to 1557 18,855 votes |
| 4 | claude-opus-4-7 | Anthropic | 1548 95% CI 1542 to 1554 17,377 votes |
| 5 | claude-opus-4-6 | Anthropic | 1547 95% CI 1541 to 1552 21,309 votes |
| 6 | kimi-k3-max | Moonshot AI | 1542 95% CI 1531 to 1553 3,111 votes |
| 7 | muse-spark-1.2 (xHigh) | Meta | 1533 95% CI 1514 to 1553 947 votes |
| 8 | claude-opus-4-8-thinking | Anthropic | 1533 95% CI 1526 to 1540 11,325 votes |
| 9 | claude-opus-5-high | Anthropic | 1531 95% CI 1522 to 1540 5,281 votes |
| 10 | muse-spark-1.1 | Meta | 1531 95% CI 1522 to 1540 4,911 votes |
グラフは表示範囲内でスケーリングされているため、小さな差異も読み取りやすくなっています。ゼロ起点ではありません。信頼区間が重複している場合、モデルは統計的に同率です。上位グループは厳密な順位ではなく、グループとして解釈してください。
この結果の読み方と、なぜ数値が異なるのか
LMArena (formerly LMSYS Chatbot Arena)
Real people are shown the same prompt answered by two anonymous models side by side and vote for the better answer. Millions of these blind head-to-head votes are fed into a Bradley-Terry statistical model (the successor to the Elo system it started with) which converts win/loss pairs into a single rating per model. A higher rating means people picked that model more often against strong opposition. This snapshot uses the 'style control' variant, which is the site's default: it statistically adjusts for answer length and formatting so a model cannot climb simply by writing longer, prettier replies.
注意すべきこと: It measures which answer people LIKE, not which answer is CORRECT - a confident, well-written wrong answer can still win a vote. Voters are self-selected volunteers rather than a representative sample, prompts skew toward what that crowd chooses to type, and models with few votes have wide confidence intervals (ci_low/ci_high) that often overlap the models ranked above and below them. Treat small rank gaps as ties.
公開済み 2026-08-11 · ライブリーダーボード ↗ · Data: LMArena leaderboard dataset (CC BY 4.0).
LiveBench
A fixed set of test questions with objectively verifiable answers is run against each model and scored automatically against ground truth - no human voting and no AI judge, so the score is repeatable. This release spans 23 tasks grouped into 7 categories. Each category score is the average of its tasks, and the headline 'global average' is the average of the 7 category scores, so every category counts equally regardless of how many tasks it contains. Scores are percentages: 100 is perfect.
注意すべきこと: Contamination-LIMITED, not contamination-proof: questions are refreshed from recent sources to reduce the chance a model simply memorised them during training, but that cannot be guaranteed. Scores reflect only these 23 tasks - they say nothing about tone, safety, speed or cost. Many entries are effort/thinking variants of the same underlying model (model_id shows the exact configuration tested), and a variant given more reasoning budget will usually outscore the cheaper default that most people actually use.
公開済み · ライブリーダーボード ↗ · Data: LiveBench 2026-06-25 release, livebench.ai.
2つの異なる単位。同一チャートには表示しない
人間の選好評価と正答率スコアは同じ軸で表せないため、このページでは両者を1つの表に混在させていません。片方でトップに立つモデルがもう片方でそうでない場合も、それは矛盾ではなく「どちらの回答が好まれたか」と「どちらの回答が正しかったか」という、本質的に異なる問いへの答えです。
このページでできないこと
どのAIを使うべきかは教えてくれません。ベンチマークのトップモデルが、あなたの用途に最適とは限らないからです。価格・可用性・インテグレーション・コンテキスト長、そしてあなたの指示にどれだけ従うかは、スコアの数点差より重要です。 どのAIを使えばいいですか？ →