All rankings / Search and documents

Search · factuality

Human preference ratings for the overall category of this Arena task. Scores from different boards are not interchangeable.

40 entries · Arena score · Higher is better

Source publication: 2026-10-13. Retrieved 2026-11-05.

Source: Arena (lmarena-ai) · CC-BY-4.0.

Ranks use confidence bounds. Several models can share a rank; the rows within a shared rank are displayed by score.

Search · factuality · Arena score
Significance rankModel / evaluated variantArena score95% confidence intervalVotes
1gpt-5.6-sol-xhighopenai1,251.131,246.28 – 1,255.9760,437
1gpt-5.5-searchopenai1,247.321,243.44 – 1,251.21125,914
1qwen3.8-maxalibaba1,240.451,228.61 – 1,252.291,870
2gpt-6-astra-maxopenai1,239.921,233.58 – 1,246.2718,483
2gpt-6.1-sol-maxopenai1,236.131,228.31 – 1,243.955,555
4claude-opus-4-6-searchanthropic1,226.821,223.6 – 1,230.04173,142
6gpt-5.2-searchopenai1,224.011,220.12 – 1,227.955,464
6claude-fable-5.1-highanthropic1,221.531,215.4 – 1,227.6712,782
6grok-4.5xai1,219.921,215.44 – 1,224.3976,458
7claude-opus-5-highanthropic1,216.861,210.71 – 1,223.0213,285
7ernie-5.1baidu1,213.761,204.71 – 1,222.814,024
8gpt-5.4-searchopenai1,212.941,209.1 – 1,216.79126,672
8claude-fable-5-highanthropic1,212.311,207.32 – 1,217.373,596
10claude-opus-4-7anthropic1,210.181,206.44 – 1,213.92127,472
10claude-opus-5.5-mediumanthropic1,205.191,198.86 – 1,211.5110,842
12gpt-5.1-searchopenai1,203.481,199.43 – 1,207.5362,869
14claude-sonnet-4-6-searchanthropic1,201.771,198.37 – 1,205.16173,930
15gpt-5.2-search-non-reasoningopenai1,197.961,194.71 – 1,201.2179,609
15gpt-6-sol-maxopenai1,197.851,191.94 – 1,203.7711,313
15grok-4.20-multi-agent-beta-0309xai1,197.771,193.6 – 1,201.94126,186
15claude-opus-4-8anthropic1,197.291,192.84 – 1,201.74106,304
15gemini-3.1-pro-groundinggoogle1,196.991,193.3 – 1,200.67151,380
15grok-4.20-beta1xai1,195.771,191.31 – 1,200.2356,876
18gemini-3-pro-groundinggoogle1,193.451,188.66 – 1,198.2438,729
24gpt-6-luna-maxopenai1,1831,177.08 – 1,188.9110,793
25claude-sonnet-5-searchanthropic1,181.751,176.92 – 1,186.5972,978
25claude-opus-4-5-searchanthropic1,180.911,176.56 – 1,185.2564,860
25gemini-3-flash-groundinggoogle1,179.131,175.72 – 1,182.54196,216
25grok-4.3xai1,174.991,171.24 – 1,178.74128,393
30claude-sonnet-4-5-searchanthropic1,162.491,158.51 – 1,166.47143,960
30gemini-2.5-pro-groundinggoogle1,162.41,158.92 – 1,165.8869,356
30ppl-sonar-reasoning-pro-highperplexity1,155.581,149.14 – 1,162.0112,927
32claude-opus-4-1-searchanthropic1,154.091,150.02 – 1,158.1762,796
32grok-4-1-fast-searchxai1,152.021,148.18 – 1,155.8685,736
32grok-4-fast-searchxai1,1461,141.23 – 1,150.7630,673
32gpt-5-searchopenai1,143.971,134.48 – 1,153.464,004
35claude-opus-4-searchanthropic1,139.61,133.53 – 1,145.6714,486
36ppl-sonar-pro-highperplexity1,132.631,126.14 – 1,139.1212,844
35grok-4-searchxai1,131.131,120.61 – 1,141.663,189
36o3-searchopenai1,129.231,119.13 – 1,139.323,818

What this ranking can tell you

  • The population is the models in this source board, including variants and models without an available API price.
  • Ranks use the published confidence bounds: one plus the number of models whose lower bound exceeds this model’s upper bound. Shared ranks are not proof of equal performance.

Check the source

The original evidence is retained byte for byte. The downloadable board includes the source URLs, revision and SHA-256 hashes.

Download this board and provenance →

Source files and licences

Explore other model types →Read the benchmark guide →

Evidence & Ask