AI leaderboard
Qual modelo lidera, no geral e por categoria. Estes são os benchmarks públicos nos quais se baseiam os gráficos que você vê no YouTube e no X. Não publicamos pontuações próprias: cada número pertence ao benchmark que o gerou, registrado com uma data e creditado à fonte.
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.
Limitações: 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.
Fonte: LMArena (formerly LMSYS Chatbot Arena) ↗ · publicado 2026-08-11 · dados subjacentes ↗ · Dataset released under Creative Commons Attribution 4.0 (CC BY 4.0). Reuse is permitted with attribution - credit LMArena and link to the leaderboard.
| # | Modelo | Organização | Arena score (human preference) |
|---|---|---|---|
| 1 | claude-fable-5 | Anthropic | 1506 95% CI 1501 to 1512 21,304 votes |
| 2 | claude-opus-4-6-thinking | Anthropic | 1505 95% CI 1501 to 1508 72,425 votes |
| 3 | claude-opus-4-7-thinking | Anthropic | 1502 95% CI 1498 to 1506 60,222 votes |
| 4 | muse-spark-1.2 (xHigh) | Meta | 1498 95% CI 1488 to 1509 3,278 votes |
| 5 | claude-opus-4-6 | Anthropic | 1498 95% CI 1494 to 1501 76,386 votes |
| 6 | claude-opus-5-high | Anthropic | 1494 95% CI 1489 to 1499 19,498 votes |
| 7 | claude-opus-4-7 | Anthropic | 1494 95% CI 1490 to 1498 61,308 votes |
| 8 | claude-opus-5-max | Anthropic | 1490 95% CI 1483 to 1497 9,419 votes |
| 9 | qwen3.8-max | Alibaba | 1490 95% CI 1482 to 1498 6,789 votes |
| 10 | muse-spark-1.1 | Meta | 1489 95% CI 1483 to 1494 16,648 votes |
| 11 | muse-spark | Meta | 1488 95% CI 1482 to 1494 13,600 votes |
| 12 | kimi-k3-max | Moonshot AI | 1487 95% CI 1481 to 1493 11,762 votes |
| 13 | gemini-3.1-pro-preview | 1486 95% CI 1483 to 1490 94,814 votes | |
| 14 | gemini-3-pro | 1486 95% CI 1482 to 1489 41,509 votes | |
| 15 | gemini-3.6-flash | 1484 95% CI 1478 to 1490 13,559 votes | |
| 16 | gpt-5.5-high | OpenAI | 1482 95% CI 1477 to 1486 55,210 votes |
| 17 | claude-opus-4-8-thinking | Anthropic | 1481 95% CI 1477 to 1486 40,409 votes |
| 18 | gpt-5.6-sol-xhigh | OpenAI | 1481 95% CI 1475 to 1487 15,304 votes |
| 19 | gemini-3.5-flash-high | 1477 95% CI 1473 to 1482 25,613 votes | |
| 20 | gpt-5.5 | OpenAI | 1477 95% CI 1473 to 1481 56,513 votes |
As barras são dimensionadas ao longo do intervalo visível para que pequenas diferenças permaneçam legíveis. Elas não começam do zero. Onde os intervalos de confiança se sobrepõem, os modelos estão estatisticamente empatados: leia o grupo do topo como um grupo, não como uma ordem estrita.
Como interpretar estes dados, e por que eles divergem
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.
Fique atento a: 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.
Publicado 2026-08-11 · placar ao vivo ↗ · 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.
Fique atento a: 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.
Publicado · placar ao vivo ↗ · Data: LiveBench 2026-06-25 release, livebench.ai.
Duas unidades diferentes, nunca um único gráfico
Uma avaliação de preferência humana e uma pontuação de acertos não podem compartilhar o mesmo eixo, por isso esta página nunca os mistura em uma única tabela. Um modelo pode liderar em um e não no outro, e isso é um sinal real sobre o que ele faz bem, não uma contradição: um pergunta "qual resposta as pessoas preferiram?", o outro pergunta "qual resposta estava correta?".
O que esta página não fará
Ele não vai dizer qual IA usar. O líder do benchmark muitas vezes não é a ferramenta certa para o seu trabalho, preço, disponibilidade, integrações, tamanho do contexto e o quanto ele segue SUAS instruções costumam importar mais do que um ou dois pontos de pontuação. Qual IA devo usar? →