LMArena Search
Arena human-vote leaderboard: rating with confidence interval, vote count and rank. One row per model and category.
No filters. Showing 42 of 42 rows.
1[1–3] AlibabaDossier | No effort set | No harness | Qwen3.8-Max, Arena rating: 1,260 (scale 950.0 to 1259.8), 95% confidence interval 1246.2 to 1273.41,260(1246.2–1273.4) | 1998 seeds | Cost kind unknown |
2[1–3] OpenAIDossierTied #1 | xhigh | No harness | GPT-5.6 Sol, Arena rating: 1,258 (scale 950.0 to 1259.8), 95% confidence interval 1251.4 to 1264.31,258(1251.4–1264.3) | 62364 seeds | Cost kind unknown |
3[2–4] AnthropicDossierTied #1 | search | No harness | Opus 4.6, Arena rating: 1,252 (scale 950.0 to 1259.8), 95% confidence interval 1247.0 to 1256.31,252(1247.0–1256.3) | 166732 seeds | Cost kind unknown |
4[3–5] OpenAIDossierTied #1 | search | No harness | GPT-5.5, Arena rating: 1,246 (scale 950.0 to 1259.8), 95% confidence interval 1240.7 to 1251.01,246(1240.7–1251.0) | 121489 seeds | Cost kind unknown |
5[4–7] AnthropicDossier | high | No harness | Opus 5, Arena rating: 1,239 (scale 950.0 to 1259.8), 95% confidence interval 1232.2 to 1245.81,239(1232.2–1245.8) | 34000 seeds | Cost kind unknown |
6[5–8] AnthropicDossier | high | No harness | Fable 5.1, Arena rating: 1,237 (scale 950.0 to 1259.8), 95% confidence interval 1230.4 to 1244.21,237(1230.4–1244.2) | 33769 seeds | Cost kind unknown |
7[5–11] OpenAIDossier | max | No harness | GPT-6 Astra, Arena rating: 1,231 (scale 950.0 to 1259.8), 95% confidence interval 1222.9 to 1238.71,231(1222.9–1238.7) | 17456 seeds | Cost kind unknown |
8[6–11] OpenAIDossier | max | No harness | GPT-6.1 Sol, Arena rating: 1,230 (scale 950.0 to 1259.8), 95% confidence interval 1222.6 to 1238.21,230(1222.6–1238.2) | 26556 seeds | Cost kind unknown |
9[6–11] AnthropicDossier | No effort set | No harness | Opus 4.7, Arena rating: 1,230 (scale 950.0 to 1259.8), 95% confidence interval 1225.1 to 1235.11,230(1225.1–1235.1) | 121890 seeds | Cost kind unknown |
10[7–12] AnthropicDossier | high | No harness | Fable 5, Arena rating: 1,228 (scale 950.0 to 1259.8), 95% confidence interval 1221.5 to 1234.81,228(1221.5–1234.8) | 69528 seeds | Cost kind unknown |
11[6–13] BaiduDossier | No effort set | No harness | ERNIE 5.1, Arena rating: 1,227 (scale 950.0 to 1259.8), 95% confidence interval 1217.2 to 1236.41,227(1217.2–1236.4) | 3925 seeds | Cost kind unknown |
12[10–13] AnthropicDossier | search | No harness | Sonnet 4.6, Arena rating: 1,221 (scale 950.0 to 1259.8), 95% confidence interval 1216.0 to 1225.21,221(1216.0–1225.2) | 167499 seeds | Cost kind unknown |
13[12–15] xAIDossier | No effort set | No harness | Grok 4.5, Arena rating: 1,216 (scale 950.0 to 1259.8), 95% confidence interval 1210.1 to 1221.61,216(1210.1–1221.6) | 105282 seeds | Cost kind unknown |
14[12–17] AnthropicDossier | medium | No harness | Opus 5.5, Arena rating: 1,213 (scale 950.0 to 1259.8), 95% confidence interval 1206.5 to 1220.11,213(1206.5–1220.1) | 31730 seeds | Cost kind unknown |
15[13–17] GoogleDossier | grounding | No harness | Gemini 3.1 Pro Preview, Arena rating: 1,210 (scale 950.0 to 1259.8), 95% confidence interval 1205.8 to 1215.01,210(1205.8–1215.0) | 171494 seeds | Cost kind unknown |
16[14–19] OpenAIDossier | search | No harness | GPT-5.2, Arena rating: 1,207 (scale 950.0 to 1259.8), 95% confidence interval 1201.5 to 1212.71,207(1201.5–1212.7) | 54658 seeds | Cost kind unknown |
17[14–19] GoogleDossier | grounding | No harness | Gemini 3 Pro Preview, Arena rating: 1,207 (scale 950.0 to 1259.8), 95% confidence interval 1201.0 to 1212.11,207(1201.0–1212.1) | 38324 seeds | Cost kind unknown |
18[15–20] xAIDossier | No effort set | No harness | Grok 4.20 Multi-Agent Beta (0309), Arena rating: 1,205 (scale 950.0 to 1259.8), 95% confidence interval 1199.6 to 1210.71,205(1199.6–1210.7) | 121069 seeds | Cost kind unknown |
19[16–21] AnthropicDossier | No effort set | No harness | Opus 4.8, Arena rating: 1,202 (scale 950.0 to 1259.8), 95% confidence interval 1196.7 to 1208.21,202(1196.7–1208.2) | 101197 seeds | Cost kind unknown |
20[18–22] OpenAIDossier | search | No harness | GPT-5.1, Arena rating: 1,199 (scale 950.0 to 1259.8), 95% confidence interval 1194.1 to 1204.71,199(1194.1–1204.7) | 61989 seeds | Cost kind unknown |
21[18–23] OpenAIDossier | search | No harness | GPT-5.4, Arena rating: 1,198 (scale 950.0 to 1259.8), 95% confidence interval 1192.5 to 1203.61,198(1192.5–1203.6) | 121580 seeds | Cost kind unknown |
22[20–23] GoogleDossier | grounding | No harness | Gemini 3 Flash Preview, Arena rating: 1,196 (scale 950.0 to 1259.8), 95% confidence interval 1191.2 to 1200.01,196(1191.2–1200.0) | 189590 seeds | Cost kind unknown |
23[21–24] xAIDossier | No effort set | No harness | Grok 4.20 Beta 1, Arena rating: 1,191 (scale 950.0 to 1259.8), 95% confidence interval 1184.8 to 1196.91,191(1184.8–1196.9) | 55902 seeds | Cost kind unknown |
24[22–25] AnthropicDossier | search | No harness | Sonnet 5, Arena rating: 1,188 (scale 950.0 to 1259.8), 95% confidence interval 1181.3 to 1194.21,188(1181.3–1194.2) | 71297 seeds | Cost kind unknown |
25[24–27] AnthropicDossier | search | No harness | Opus 4.5, Arena rating: 1,179 (scale 950.0 to 1259.8), 95% confidence interval 1173.5 to 1184.81,179(1173.5–1184.8) | 63904 seeds | Cost kind unknown |
26[25–30] OpenAIDossier | max | No harness | GPT-6 Sol, Arena rating: 1,174 (scale 950.0 to 1259.8), 95% confidence interval 1167.0 to 1181.91,174(1167.0–1181.9) | 13499 seeds | Cost kind unknown |
27[26–30] OpenAIDossier | search · non-thinking | No harness | GPT-5.2, Arena rating: 1,172 (scale 950.0 to 1259.8), 95% confidence interval 1166.3 to 1177.51,172(1166.3–1177.5) | 78443 seeds | Cost kind unknown |
28[26–30] xAIDossier | search | No harness | Grok 4.1 Fast, Arena rating: 1,172 (scale 950.0 to 1259.8), 95% confidence interval 1166.6 to 1176.61,172(1166.6–1176.6) | 84455 seeds | Cost kind unknown |
29[26–30] xAIDossier | search | No harness | Grok 4 Fast, Arena rating: 1,171 (scale 950.0 to 1259.8), 95% confidence interval 1165.9 to 1175.41,171(1165.9–1175.4) | 42640 seeds | Cost kind unknown |
30[27–31] xAIDossier | No effort set | No harness | Grok 4.3, Arena rating: 1,166 (scale 950.0 to 1259.8), 95% confidence interval 1161.6 to 1171.31,166(1161.6–1171.3) | 122703 seeds | Cost kind unknown |
31[30–32] OpenAIDossier | max | No harness | GPT-6 Luna, Arena rating: 1,160 (scale 950.0 to 1259.8), 95% confidence interval 1153.2 to 1167.11,160(1153.2–1167.1) | 46776 seeds | Cost kind unknown |
32[31–32] AnthropicDossier | search | No harness | Sonnet 4.5, Arena rating: 1,159 (scale 950.0 to 1259.8), 95% confidence interval 1154.0 to 1164.41,159(1154.0–1164.4) | 139468 seeds | Cost kind unknown |
33[33–35] AnthropicDossier | search | No harness | Opus 4.1, Arena rating: 1,148 (scale 950.0 to 1259.8), 95% confidence interval 1143.2 to 1153.11,148(1143.2–1153.1) | 78777 seeds | Cost kind unknown |
34[33–37] OpenAIDossier | search | No harness | o3, Arena rating: 1,145 (scale 950.0 to 1259.8), 95% confidence interval 1138.9 to 1150.21,145(1138.9–1150.2) | 20655 seeds | Cost kind unknown |
35[33–37] GoogleDossier | grounding | No harness | Gemini 2.5 Pro, Arena rating: 1,143 (scale 950.0 to 1259.8), 95% confidence interval 1138.6 to 1148.21,143(1138.6–1148.2) | 85505 seeds | Cost kind unknown |
36[33–37] xAIDossier | search | No harness | Grok 4, Arena rating: 1,142 (scale 950.0 to 1259.8), 95% confidence interval 1135.7 to 1148.01,142(1135.7–1148.0) | 19069 seeds | Cost kind unknown |
37[34–38] PerplexityDossier | high | No harness | Sonar Reasoning Pro, Arena rating: 1,139 (scale 950.0 to 1259.8), 95% confidence interval 1132.9 to 1144.71,139(1132.9–1144.7) | 29197 seeds | Cost kind unknown |
38[37–40] OpenAIDossier | search | No harness | GPT-5, Arena rating: 1,133 (scale 950.0 to 1259.8), 95% confidence interval 1127.0 to 1139.11,133(1127.0–1139.1) | 20742 seeds | Cost kind unknown |
39[38–40] PerplexityDossier | high | No harness | Sonar Pro, Arena rating: 1,130 (scale 950.0 to 1259.8), 95% confidence interval 1123.8 to 1135.61,130(1123.8–1135.6) | 28684 seeds | Cost kind unknown |
40[38–40] AnthropicDossier | search | No harness | Opus 4, Arena rating: 1,127 (scale 950.0 to 1259.8), 95% confidence interval 1121.5 to 1132.51,127(1121.5–1132.5) | 31275 seeds | Cost kind unknown |
41 DiffbotDossier | No effort set | No harness | Diffbot Small XL, Arena rating: 1,022 (scale 950.0 to 1259.8), 95% confidence interval 1013.5 to 1030.61,022(1013.5–1030.6) | 6339 seeds | Cost kind unknown |
42 OpenAIDossier | search | No harness | GPT-4o, Arena rating: 1,006 (scale 950.0 to 1259.8), 95% confidence interval 995.1 to 1017.61,006(995.1–1017.6) | 3392 seeds | Cost kind unknown |
- 1AlibabaDossier
Arena rating
Qwen3.8-Max, Arena rating: 1,260 (scale 950.0 to 1259.8), 95% confidence interval 1246.2 to 1273.41,260(1246.2–1273.4) Arena rating
GPT-5.6 Sol, Arena rating: 1,258 (scale 950.0 to 1259.8), 95% confidence interval 1251.4 to 1264.31,258(1251.4–1264.3)Arena rating
Opus 4.6, Arena rating: 1,252 (scale 950.0 to 1259.8), 95% confidence interval 1247.0 to 1256.31,252(1247.0–1256.3)Arena rating
GPT-5.5, Arena rating: 1,246 (scale 950.0 to 1259.8), 95% confidence interval 1240.7 to 1251.01,246(1240.7–1251.0)- 5AnthropicDossier
high
Arena rating
Opus 5, Arena rating: 1,239 (scale 950.0 to 1259.8), 95% confidence interval 1232.2 to 1245.81,239(1232.2–1245.8) - 6AnthropicDossier
high
Arena rating
Fable 5.1, Arena rating: 1,237 (scale 950.0 to 1259.8), 95% confidence interval 1230.4 to 1244.21,237(1230.4–1244.2) - 7OpenAIDossier
max
Arena rating
GPT-6 Astra, Arena rating: 1,231 (scale 950.0 to 1259.8), 95% confidence interval 1222.9 to 1238.71,231(1222.9–1238.7) - 8OpenAIDossier
max
Arena rating
GPT-6.1 Sol, Arena rating: 1,230 (scale 950.0 to 1259.8), 95% confidence interval 1222.6 to 1238.21,230(1222.6–1238.2) - 9AnthropicDossier
Arena rating
Opus 4.7, Arena rating: 1,230 (scale 950.0 to 1259.8), 95% confidence interval 1225.1 to 1235.11,230(1225.1–1235.1) - 10AnthropicDossier
high
Arena rating
Fable 5, Arena rating: 1,228 (scale 950.0 to 1259.8), 95% confidence interval 1221.5 to 1234.81,228(1221.5–1234.8) - 11BaiduDossier
Arena rating
ERNIE 5.1, Arena rating: 1,227 (scale 950.0 to 1259.8), 95% confidence interval 1217.2 to 1236.41,227(1217.2–1236.4) - 12AnthropicDossier
search
Arena rating
Sonnet 4.6, Arena rating: 1,221 (scale 950.0 to 1259.8), 95% confidence interval 1216.0 to 1225.21,221(1216.0–1225.2) - 13xAIDossier
Arena rating
Grok 4.5, Arena rating: 1,216 (scale 950.0 to 1259.8), 95% confidence interval 1210.1 to 1221.61,216(1210.1–1221.6) - 14AnthropicDossier
medium
Arena rating
Opus 5.5, Arena rating: 1,213 (scale 950.0 to 1259.8), 95% confidence interval 1206.5 to 1220.11,213(1206.5–1220.1) - 15GoogleDossier
grounding
Arena rating
Gemini 3.1 Pro Preview, Arena rating: 1,210 (scale 950.0 to 1259.8), 95% confidence interval 1205.8 to 1215.01,210(1205.8–1215.0) - 16OpenAIDossier
search
Arena rating
GPT-5.2, Arena rating: 1,207 (scale 950.0 to 1259.8), 95% confidence interval 1201.5 to 1212.71,207(1201.5–1212.7) - 17GoogleDossier
grounding
Arena rating
Gemini 3 Pro Preview, Arena rating: 1,207 (scale 950.0 to 1259.8), 95% confidence interval 1201.0 to 1212.11,207(1201.0–1212.1) - 18xAIDossier
Arena rating
Grok 4.20 Multi-Agent Beta (0309), Arena rating: 1,205 (scale 950.0 to 1259.8), 95% confidence interval 1199.6 to 1210.71,205(1199.6–1210.7) - 19AnthropicDossier
Arena rating
Opus 4.8, Arena rating: 1,202 (scale 950.0 to 1259.8), 95% confidence interval 1196.7 to 1208.21,202(1196.7–1208.2) - 20OpenAIDossier
search
Arena rating
GPT-5.1, Arena rating: 1,199 (scale 950.0 to 1259.8), 95% confidence interval 1194.1 to 1204.71,199(1194.1–1204.7) - 21OpenAIDossier
search
Arena rating
GPT-5.4, Arena rating: 1,198 (scale 950.0 to 1259.8), 95% confidence interval 1192.5 to 1203.61,198(1192.5–1203.6) - 22GoogleDossier
grounding
Arena rating
Gemini 3 Flash Preview, Arena rating: 1,196 (scale 950.0 to 1259.8), 95% confidence interval 1191.2 to 1200.01,196(1191.2–1200.0) - 23xAIDossier
Arena rating
Grok 4.20 Beta 1, Arena rating: 1,191 (scale 950.0 to 1259.8), 95% confidence interval 1184.8 to 1196.91,191(1184.8–1196.9) - 24AnthropicDossier
search
Arena rating
Sonnet 5, Arena rating: 1,188 (scale 950.0 to 1259.8), 95% confidence interval 1181.3 to 1194.21,188(1181.3–1194.2) - 25AnthropicDossier
search
Arena rating
Opus 4.5, Arena rating: 1,179 (scale 950.0 to 1259.8), 95% confidence interval 1173.5 to 1184.81,179(1173.5–1184.8) - 26OpenAIDossier
max
Arena rating
GPT-6 Sol, Arena rating: 1,174 (scale 950.0 to 1259.8), 95% confidence interval 1167.0 to 1181.91,174(1167.0–1181.9) - 27OpenAIDossier
search · non-thinking
Arena rating
GPT-5.2, Arena rating: 1,172 (scale 950.0 to 1259.8), 95% confidence interval 1166.3 to 1177.51,172(1166.3–1177.5) - 28xAIDossier
search
Arena rating
Grok 4.1 Fast, Arena rating: 1,172 (scale 950.0 to 1259.8), 95% confidence interval 1166.6 to 1176.61,172(1166.6–1176.6) - 29xAIDossier
search
Arena rating
Grok 4 Fast, Arena rating: 1,171 (scale 950.0 to 1259.8), 95% confidence interval 1165.9 to 1175.41,171(1165.9–1175.4) - 30xAIDossier
Arena rating
Grok 4.3, Arena rating: 1,166 (scale 950.0 to 1259.8), 95% confidence interval 1161.6 to 1171.31,166(1161.6–1171.3) - 31OpenAIDossier
max
Arena rating
GPT-6 Luna, Arena rating: 1,160 (scale 950.0 to 1259.8), 95% confidence interval 1153.2 to 1167.11,160(1153.2–1167.1) - 32AnthropicDossier
search
Arena rating
Sonnet 4.5, Arena rating: 1,159 (scale 950.0 to 1259.8), 95% confidence interval 1154.0 to 1164.41,159(1154.0–1164.4) - 33AnthropicDossier
search
Arena rating
Opus 4.1, Arena rating: 1,148 (scale 950.0 to 1259.8), 95% confidence interval 1143.2 to 1153.11,148(1143.2–1153.1) - 34OpenAIDossier
search
Arena rating
o3, Arena rating: 1,145 (scale 950.0 to 1259.8), 95% confidence interval 1138.9 to 1150.21,145(1138.9–1150.2) - 35GoogleDossier
grounding
Arena rating
Gemini 2.5 Pro, Arena rating: 1,143 (scale 950.0 to 1259.8), 95% confidence interval 1138.6 to 1148.21,143(1138.6–1148.2) - 36xAIDossier
search
Arena rating
Grok 4, Arena rating: 1,142 (scale 950.0 to 1259.8), 95% confidence interval 1135.7 to 1148.01,142(1135.7–1148.0) - 37PerplexityDossier
high
Arena rating
Sonar Reasoning Pro, Arena rating: 1,139 (scale 950.0 to 1259.8), 95% confidence interval 1132.9 to 1144.71,139(1132.9–1144.7) - 38OpenAIDossier
search
Arena rating
GPT-5, Arena rating: 1,133 (scale 950.0 to 1259.8), 95% confidence interval 1127.0 to 1139.11,133(1127.0–1139.1) - 39PerplexityDossier
high
Arena rating
Sonar Pro, Arena rating: 1,130 (scale 950.0 to 1259.8), 95% confidence interval 1123.8 to 1135.61,130(1123.8–1135.6) - 40AnthropicDossier
search
Arena rating
Opus 4, Arena rating: 1,127 (scale 950.0 to 1259.8), 95% confidence interval 1121.5 to 1132.51,127(1121.5–1132.5) - 41DiffbotDossier
Arena rating
Diffbot Small XL, Arena rating: 1,022 (scale 950.0 to 1259.8), 95% confidence interval 1013.5 to 1030.61,022(1013.5–1030.6) - 42OpenAIDossier
search
Arena rating
GPT-4o, Arena rating: 1,006 (scale 950.0 to 1259.8), 95% confidence interval 995.1 to 1017.61,006(995.1–1017.6)