Acropolis

LMArena Vision (style control)

Arena human-vote leaderboard: rating with confidence interval, vote count and rank. One row per model and category.

Vision
Independent. Run by an operator independent of the model vendors.LMArena by ArenaVersion 1Data as of 17 Nov 2026

No filters. Showing 167 of 167 rows.

  1. 1
    AnthropicDossier

    high

    Arena rating

    Fable 5, Arena rating: 1,309 (scale 800.0 to 1309.4), 95% confidence interval 1302.5 to 1316.31,309(1302.5–1316.3)
  2. 2
    AnthropicDossier
    Tied #1

    Arena rating

    Opus 4.7, Arena rating: 1,300 (scale 800.0 to 1309.4), 95% confidence interval 1293.7 to 1306.81,300(1293.7–1306.8)
  3. 3
    AlibabaDossier

    Arena rating

    Qwen3.8-Max, Arena rating: 1,299 (scale 800.0 to 1309.4), 95% confidence interval 1291.6 to 1306.51,299(1291.6–1306.5)
  4. 4
    AnthropicDossier

    high

    Arena rating

    Opus 4.6, Arena rating: 1,299 (scale 800.0 to 1309.4), 95% confidence interval 1292.4 to 1305.11,299(1292.4–1305.1)
  5. 5
    AnthropicDossier

    high

    Arena rating

    Opus 4.7, Arena rating: 1,298 (scale 800.0 to 1309.4), 95% confidence interval 1291.9 to 1305.01,298(1291.9–1305.0)
  6. 6
    AnthropicDossier

    high

    Arena rating

    Opus 5.5, Arena rating: 1,295 (scale 800.0 to 1309.4), 95% confidence interval 1285.4 to 1305.21,295(1285.4–1305.2)
  7. 7

    Arena rating

    Muse Spark, Arena rating: 1,294 (scale 800.0 to 1309.4), 95% confidence interval 1285.3 to 1303.51,294(1285.3–1303.5)
  8. 8
    GoogleDossier

    high

    Arena rating

    Gemini 3.7 Flash, Arena rating: 1,294 (scale 800.0 to 1309.4), 95% confidence interval 1285.2 to 1302.31,294(1285.2–1302.3)
  9. 9

    xhigh

    Tied #1

    Arena rating

    Muse Spark 1.2, Arena rating: 1,294 (scale 800.0 to 1309.4), 95% confidence interval 1279.1 to 1307.91,294(1279.1–1307.9)
  10. 10
    AnthropicDossier

    Arena rating

    Opus 4.6, Arena rating: 1,293 (scale 800.0 to 1309.4), 95% confidence interval 1287.0 to 1299.51,293(1287.0–1299.5)
  11. 11
    GoogleDossier

    high

    Arena rating

    Gemini 3.8 Flash, Arena rating: 1,293 (scale 800.0 to 1309.4), 95% confidence interval 1284.9 to 1301.31,293(1284.9–1301.3)
  12. 12
    OpenAIDossier

    max

    Arena rating

    GPT-6.1 Sol, Arena rating: 1,293 (scale 800.0 to 1309.4), 95% confidence interval 1282.0 to 1303.01,293(1282.0–1303.0)
  13. 13

    max

    Arena rating

    Muse Spark 1.3, Arena rating: 1,291 (scale 800.0 to 1309.4), 95% confidence interval 1283.4 to 1299.61,291(1283.4–1299.6)
  14. 14
    GoogleDossier

    Arena rating

    Gemini 3 Pro Preview, Arena rating: 1,289 (scale 800.0 to 1309.4), 95% confidence interval 1282.0 to 1296.81,289(1282.0–1296.8)
  15. 15
    AnthropicDossier

    max

    Arena rating

    Fable 5.1, Arena rating: 1,289 (scale 800.0 to 1309.4), 95% confidence interval 1280.7 to 1297.91,289(1280.7–1297.9)
  16. 16
    OpenAIDossier

    xhigh

    Arena rating

    GPT-5.6 Sol, Arena rating: 1,288 (scale 800.0 to 1309.4), 95% confidence interval 1280.8 to 1295.21,288(1280.8–1295.2)
  17. 17

    xhigh

    Arena rating

    Muse Spark 1.3, Arena rating: 1,288 (scale 800.0 to 1309.4), 95% confidence interval 1279.7 to 1296.21,288(1279.7–1296.2)
  18. 18
    AnthropicDossier

    high

    Arena rating

    Opus 4.8, Arena rating: 1,286 (scale 800.0 to 1309.4), 95% confidence interval 1279.5 to 1292.41,286(1279.5–1292.4)
  19. 19
    AnthropicDossier

    high

    Arena rating

    Opus 5, Arena rating: 1,285 (scale 800.0 to 1309.4), 95% confidence interval 1278.9 to 1291.91,285(1278.9–1291.9)
  20. 20
    OpenAIDossier

    Arena rating

    GPT-5.5, Arena rating: 1,285 (scale 800.0 to 1309.4), 95% confidence interval 1279.0 to 1290.31,285(1279.0–1290.3)
  21. 21
    OpenAIDossier

    max

    Arena rating

    GPT-6 Astra, Arena rating: 1,283 (scale 800.0 to 1309.4), 95% confidence interval 1274.2 to 1292.61,283(1274.2–1292.6)
  22. 22
    OpenAIDossier

    high

    Arena rating

    GPT-5.5, Arena rating: 1,283 (scale 800.0 to 1309.4), 95% confidence interval 1277.6 to 1289.01,283(1277.6–1289.0)
  23. 23
    OpenAIDossier

    high

    Arena rating

    GPT-5.4, Arena rating: 1,282 (scale 800.0 to 1309.4), 95% confidence interval 1276.6 to 1287.91,282(1276.6–1287.9)
  24. 24
    GoogleDossier

    medium

    Arena rating

    Gemini 3.5 Flash, Arena rating: 1,281 (scale 800.0 to 1309.4), 95% confidence interval 1274.9 to 1288.11,281(1274.9–1288.1)
  25. 25
    GoogleDossier

    high

    Arena rating

    Gemini 3.5 Flash, Arena rating: 1,281 (scale 800.0 to 1309.4), 95% confidence interval 1274.4 to 1287.71,281(1274.4–1287.7)
  26. 26

    Arena rating

    Grok 4.5, Arena rating: 1,280 (scale 800.0 to 1309.4), 95% confidence interval 1273.6 to 1287.31,280(1273.6–1287.3)
  27. 27
    GoogleDossier

    Arena rating

    Gemini 3.1 Pro Preview, Arena rating: 1,280 (scale 800.0 to 1309.4), 95% confidence interval 1274.9 to 1284.91,280(1274.9–1284.9)
  28. 28
    GoogleDossier

    high

    Arena rating

    Gemini 3.6 Flash, Arena rating: 1,280 (scale 800.0 to 1309.4), 95% confidence interval 1272.3 to 1287.41,280(1272.3–1287.4)
  29. 29

    Arena rating

    Muse Spark 1.1, Arena rating: 1,279 (scale 800.0 to 1309.4), 95% confidence interval 1272.5 to 1286.21,279(1272.5–1286.2)
  30. 30
    AnthropicDossier

    Arena rating

    Opus 4.8, Arena rating: 1,279 (scale 800.0 to 1309.4), 95% confidence interval 1272.3 to 1285.21,279(1272.3–1285.2)
  31. 31
    OpenAIDossier

    Arena rating

    GPT-5.4, Arena rating: 1,278 (scale 800.0 to 1309.4), 95% confidence interval 1271.6 to 1284.51,278(1271.6–1284.5)
  32. 32
    OpenAIDossier

    Arena rating

    GPT-5.2 Chat, Arena rating: 1,278 (scale 800.0 to 1309.4), 95% confidence interval 1271.2 to 1284.21,278(1271.2–1284.2)
  33. 33
    OpenAIDossier

    Arena rating

    GPT-5.5 Instant, Arena rating: 1,277 (scale 800.0 to 1309.4), 95% confidence interval 1268.7 to 1286.11,277(1268.7–1286.1)
  34. 34
    AnthropicDossier

    Arena rating

    Sonnet 4.6, Arena rating: 1,275 (scale 800.0 to 1309.4), 95% confidence interval 1268.8 to 1281.01,275(1268.8–1281.0)
  35. 35
    GoogleDossier

    Arena rating

    Gemini 3 Flash Preview, Arena rating: 1,272 (scale 800.0 to 1309.4), 95% confidence interval 1267.0 to 1277.41,272(1267.0–1277.4)
  36. 36

    Arena rating

    GLM-5.3 Flash, Arena rating: 1,272 (scale 800.0 to 1309.4), 95% confidence interval 1266.3 to 1277.41,272(1266.3–1277.4)
  37. 37
    OpenAIDossier

    xhigh

    Arena rating

    GPT-5.6 Terra, Arena rating: 1,266 (scale 800.0 to 1309.4), 95% confidence interval 1259.4 to 1273.41,266(1259.4–1273.4)
  38. 38
    GoogleDossier

    Arena rating

    Gemini 3.5 Flash-Lite, Arena rating: 1,266 (scale 800.0 to 1309.4), 95% confidence interval 1258.3 to 1273.31,266(1258.3–1273.3)
  39. 39
    GoogleDossier

    Arena rating

    Gemma 4 31B, Arena rating: 1,265 (scale 800.0 to 1309.4), 95% confidence interval 1259.3 to 1270.41,265(1259.3–1270.4)
  40. 40
    Moonshot AIDossier

    Arena rating

    Kimi K2.6, Arena rating: 1,265 (scale 800.0 to 1309.4), 95% confidence interval 1258.0 to 1271.51,265(1258.0–1271.5)
  41. 41

    high

    Arena rating

    Grok 4.6, Arena rating: 1,263 (scale 800.0 to 1309.4), 95% confidence interval 1255.3 to 1271.11,263(1255.3–1271.1)
  42. 42
    DeepSeekDossier

    max

    Arena rating

    DeepSeek V4.1 Flash, Arena rating: 1,263 (scale 800.0 to 1309.4), 95% confidence interval 1252.7 to 1273.01,263(1252.7–1273.0)
  43. 43
    AnthropicDossier

    xhigh

    Arena rating

    Sonnet 5.5, Arena rating: 1,263 (scale 800.0 to 1309.4), 95% confidence interval 1252.5 to 1272.91,263(1252.5–1272.9)
  44. 44
    AlibabaDossier

    Arena rating

    Qwen3.7 Plus, Arena rating: 1,262 (scale 800.0 to 1309.4), 95% confidence interval 1255.9 to 1268.91,262(1255.9–1268.9)
  45. 45
    AnthropicDossier

    high

    Arena rating

    Sonnet 5, Arena rating: 1,261 (scale 800.0 to 1309.4), 95% confidence interval 1254.1 to 1267.51,261(1254.1–1267.5)
  46. 46
    GoogleDossier

    minimal

    Arena rating

    Gemini 3 Flash Preview, Arena rating: 1,260 (scale 800.0 to 1309.4), 95% confidence interval 1254.3 to 1264.91,260(1254.3–1264.9)
  47. 47
    OpenAIDossier

    max

    Arena rating

    GPT-6 Sol, Arena rating: 1,258 (scale 800.0 to 1309.4), 95% confidence interval 1248.4 to 1267.81,258(1248.4–1267.8)
  48. 48
    ByteDanceDossier

    Arena rating

    Dola Seed 2.0 Pro, Arena rating: 1,256 (scale 800.0 to 1309.4), 95% confidence interval 1249.1 to 1263.51,256(1249.1–1263.5)
  49. 49

    thinking

    Arena rating

    Grok 4.20 0309, Arena rating: 1,256 (scale 800.0 to 1309.4), 95% confidence interval 1249.6 to 1261.71,256(1249.6–1261.7)
  50. 50
    OpenAIDossier

    xhigh

    Arena rating

    GPT-5.6 Luna, Arena rating: 1,255 (scale 800.0 to 1309.4), 95% confidence interval 1248.0 to 1262.81,255(1248.0–1262.8)
  51. 51
    Moonshot AIDossier

    thinking

    Arena rating

    Kimi K2.5, Arena rating: 1,252 (scale 800.0 to 1309.4), 95% confidence interval 1246.5 to 1257.71,252(1246.5–1257.7)
  52. 52
    OpenAIDossier

    high

    Arena rating

    GPT-5.1, Arena rating: 1,251 (scale 800.0 to 1309.4), 95% confidence interval 1243.2 to 1258.81,251(1243.2–1258.8)
  53. 53

    Arena rating

    Grok 4.20 Multi-Agent Beta (0309), Arena rating: 1,251 (scale 800.0 to 1309.4), 95% confidence interval 1244.5 to 1256.61,251(1244.5–1256.6)
  54. 54
    OpenAIDossier

    high

    Arena rating

    GPT-5.4 mini, Arena rating: 1,250 (scale 800.0 to 1309.4), 95% confidence interval 1243.8 to 1256.11,250(1243.8–1256.1)
  55. 55

    xhigh

    Arena rating

    Grok 4.7, Arena rating: 1,249 (scale 800.0 to 1309.4), 95% confidence interval 1240.1 to 1257.81,249(1240.1–1257.8)
  56. 56
    XiaomiDossier

    Arena rating

    MiMo-V2.6-Flash, Arena rating: 1,248 (scale 800.0 to 1309.4), 95% confidence interval 1239.3 to 1256.51,248(1239.3–1256.5)
  57. 57
    GoogleDossier

    Arena rating

    Gemini 2.5 Pro, Arena rating: 1,247 (scale 800.0 to 1309.4), 95% confidence interval 1243.0 to 1252.01,247(1243.0–1252.0)
  58. 58
    XiaomiDossier

    Arena rating

    MiMo-V2.6-Pro, Arena rating: 1,247 (scale 800.0 to 1309.4), 95% confidence interval 1232.8 to 1262.11,247(1232.8–1262.1)
  59. 59
    AlibabaDossier

    Arena rating

    Qwen3.5 397B A17B, Arena rating: 1,246 (scale 800.0 to 1309.4), 95% confidence interval 1240.5 to 1250.91,246(1240.5–1250.9)
  60. 60
    StepFunDossier

    Arena rating

    Step 5 Preview, Arena rating: 1,245 (scale 800.0 to 1309.4), 95% confidence interval 1234.5 to 1256.41,245(1234.5–1256.4)
  61. 61
    AlibabaDossier

    Arena rating

    Qwen3.8-27B, Arena rating: 1,243 (scale 800.0 to 1309.4), 95% confidence interval 1235.4 to 1250.61,243(1235.4–1250.6)
  62. 62
    OpenAIDossier

    high

    Arena rating

    GPT-5.2, Arena rating: 1,242 (scale 800.0 to 1309.4), 95% confidence interval 1236.4 to 1248.51,242(1236.4–1248.5)
  63. 63
    GoogleDossier

    Arena rating

    Gemma 4 26B A4B, Arena rating: 1,242 (scale 800.0 to 1309.4), 95% confidence interval 1235.4 to 1248.21,242(1235.4–1248.2)
  64. 64

    Arena rating

    Grok 4.3, Arena rating: 1,242 (scale 800.0 to 1309.4), 95% confidence interval 1235.9 to 1247.41,242(1235.9–1247.4)
  65. 65
    OpenAIDossier

    Arena rating

    GPT-5.1, Arena rating: 1,240 (scale 800.0 to 1309.4), 95% confidence interval 1232.6 to 1247.61,240(1232.6–1247.6)
  66. 66
    OpenAIDossier

    Arena rating

    ChatGPT-4o Latest, Arena rating: 1,238 (scale 800.0 to 1309.4), 95% confidence interval 1233.0 to 1244.01,238(1233.0–1244.0)
  67. 67
    Moonshot AIDossier

    instant

    Arena rating

    Kimi K2.5, Arena rating: 1,236 (scale 800.0 to 1309.4), 95% confidence interval 1224.8 to 1246.21,236(1224.8–1246.2)
  68. 68
    OpenAIDossier

    max

    Arena rating

    GPT-6 Luna, Arena rating: 1,235 (scale 800.0 to 1309.4), 95% confidence interval 1226.4 to 1244.31,235(1226.4–1244.3)
  69. 69
    GoogleDossier

    Arena rating

    Gemini 3.1 Flash-Lite Preview, Arena rating: 1,235 (scale 800.0 to 1309.4), 95% confidence interval 1229.0 to 1240.41,235(1229.0–1240.4)
  70. 70
    XiaomiDossier

    Arena rating

    MiMo-V2.5, Arena rating: 1,234 (scale 800.0 to 1309.4), 95% confidence interval 1228.3 to 1239.91,234(1228.3–1239.9)
  71. 71
    MiniMaxDossier

    Arena rating

    MiniMax M3, Arena rating: 1,232 (scale 800.0 to 1309.4), 95% confidence interval 1225.6 to 1238.11,232(1225.6–1238.1)
  72. 72

    Arena rating

    GLM-5V-Turbo, Arena rating: 1,232 (scale 800.0 to 1309.4), 95% confidence interval 1226.3 to 1237.11,232(1226.3–1237.1)
  73. 73
    MistralDossier

    Arena rating

    Mistral Large 4, Arena rating: 1,231 (scale 800.0 to 1309.4), 95% confidence interval 1216.7 to 1244.91,231(1216.7–1244.9)
  74. 74
    GoogleDossier

    Arena rating

    Gemini 2.5 Flash Preview (09-2025), Arena rating: 1,229 (scale 800.0 to 1309.4), 95% confidence interval 1219.3 to 1239.01,229(1219.3–1239.0)
  75. 75
    OpenAIDossier

    Arena rating

    GPT-5.2, Arena rating: 1,228 (scale 800.0 to 1309.4), 95% confidence interval 1222.4 to 1234.31,228(1222.4–1234.3)
  76. 76
    AlibabaDossier

    Arena rating

    Qwen3.5 122B A10B, Arena rating: 1,226 (scale 800.0 to 1309.4), 95% confidence interval 1219.1 to 1232.61,226(1219.1–1232.6)
  77. 77
    OpenAIDossier

    Arena rating

    GPT-5 Chat, Arena rating: 1,225 (scale 800.0 to 1309.4), 95% confidence interval 1218.3 to 1232.41,225(1218.3–1232.4)
  78. 78
    OpenAIDossier

    Arena rating

    GPT-4.5 Preview, Arena rating: 1,225 (scale 800.0 to 1309.4), 95% confidence interval 1213.6 to 1237.01,225(1213.6–1237.0)
  79. 79
    BaiduDossier

    Arena rating

    ERNIE 5.0 Preview (1220), Arena rating: 1,219 (scale 800.0 to 1309.4), 95% confidence interval 1207.9 to 1230.31,219(1207.9–1230.3)
  80. 80
    XiaomiDossier

    Arena rating

    MiMo-V2-Omni, Arena rating: 1,216 (scale 800.0 to 1309.4), 95% confidence interval 1208.9 to 1223.51,216(1208.9–1223.5)
  81. 81
    AlibabaDossier

    Arena rating

    Qwen3.5 27B, Arena rating: 1,216 (scale 800.0 to 1309.4), 95% confidence interval 1210.3 to 1221.11,216(1210.3–1221.1)
  82. 82
    OpenAIDossier

    Arena rating

    o3, Arena rating: 1,216 (scale 800.0 to 1309.4), 95% confidence interval 1209.0 to 1222.11,216(1209.0–1222.1)
  83. 83
    GoogleDossier

    Arena rating

    Gemini 2.5 Flash, Arena rating: 1,215 (scale 800.0 to 1309.4), 95% confidence interval 1210.7 to 1219.81,215(1210.7–1219.8)
  84. 84
    AlibabaDossier

    Arena rating

    Qwen3 VL 235B A22B Instruct, Arena rating: 1,214 (scale 800.0 to 1309.4), 95% confidence interval 1207.5 to 1221.51,214(1207.5–1221.5)
  85. 85
    OpenAIDossier

    Arena rating

    GPT-4.1, Arena rating: 1,214 (scale 800.0 to 1309.4), 95% confidence interval 1207.0 to 1220.41,214(1207.0–1220.4)
  86. 86
    NVIDIADossier

    Arena rating

    Nemotron 3.5 Super VL 120B A12B (NVFP4), Arena rating: 1,212 (scale 800.0 to 1309.4), 95% confidence interval 1202.8 to 1221.41,212(1202.8–1221.4)
  87. 87
    OpenAIDossier

    high

    Arena rating

    GPT-5, Arena rating: 1,211 (scale 800.0 to 1309.4), 95% confidence interval 1204.0 to 1218.11,211(1204.0–1218.1)
  88. 88
    AnthropicDossier

    thinking · 16,000 budget

    Arena rating

    Opus 4, Arena rating: 1,208 (scale 800.0 to 1309.4), 95% confidence interval 1191.8 to 1223.91,208(1191.8–1223.9)
  89. 89
    Thinking MachinesDossier

    Arena rating

    Inkling Small, Arena rating: 1,208 (scale 800.0 to 1309.4), 95% confidence interval 1199.8 to 1215.51,208(1199.8–1215.5)
  90. 90
    AnthropicDossier

    thinking · 32,000 budget

    Arena rating

    Claude Sonnet 4, Arena rating: 1,206 (scale 800.0 to 1309.4), 95% confidence interval 1189.2 to 1223.11,206(1189.2–1223.1)
  91. 91
    OpenAIDossier

    Arena rating

    o4-mini, Arena rating: 1,201 (scale 800.0 to 1309.4), 95% confidence interval 1194.1 to 1208.41,201(1194.1–1208.4)
  92. 92
    OpenAIDossier

    Arena rating

    GPT-4.1 mini, Arena rating: 1,201 (scale 800.0 to 1309.4), 95% confidence interval 1193.3 to 1208.41,201(1193.3–1208.4)
  93. 93
    MistralDossier

    Arena rating

    Mistral Large 3, Arena rating: 1,200 (scale 800.0 to 1309.4), 95% confidence interval 1192.7 to 1207.91,200(1192.7–1207.9)
  94. 94
    MistralDossier

    Arena rating

    Mistral Medium 3.5, Arena rating: 1,199 (scale 800.0 to 1309.4), 95% confidence interval 1189.6 to 1208.81,199(1189.6–1208.8)
  95. 95
    OpenAIDossier

    high

    Arena rating

    GPT-5.4 nano, Arena rating: 1,199 (scale 800.0 to 1309.4), 95% confidence interval 1192.5 to 1205.11,199(1192.5–1205.1)
  96. 96

    thinking

    Arena rating

    Grok 4.1 Fast, Arena rating: 1,196 (scale 800.0 to 1309.4), 95% confidence interval 1189.6 to 1203.31,196(1189.6–1203.3)
  97. 97
    OpenAIDossier

    Arena rating

    o1, Arena rating: 1,193 (scale 800.0 to 1309.4), 95% confidence interval 1182.1 to 1203.61,193(1182.1–1203.6)
  98. 98
    AnthropicDossier

    thinking · 32,000 budget

    Arena rating

    Claude 3.7 Sonnet, Arena rating: 1,193 (scale 800.0 to 1309.4), 95% confidence interval 1177.1 to 1208.21,193(1177.1–1208.2)
  99. 99
    AnthropicDossier

    Arena rating

    Opus 4, Arena rating: 1,189 (scale 800.0 to 1309.4), 95% confidence interval 1176.0 to 1201.81,189(1176.0–1201.8)
  100. 100
    GoogleDossier

    thinking

    Arena rating

    Gemini 2.5 Flash-Lite Preview (06-17), Arena rating: 1,189 (scale 800.0 to 1309.4), 95% confidence interval 1181.3 to 1196.21,189(1181.3–1196.2)
  101. 101
    AnthropicDossier

    Arena rating

    Claude Sonnet 4, Arena rating: 1,189 (scale 800.0 to 1309.4), 95% confidence interval 1174.5 to 1202.61,189(1174.5–1202.6)
  102. 102
    AlibabaDossier

    Arena rating

    Qwen3 VL 235B A22B Thinking, Arena rating: 1,188 (scale 800.0 to 1309.4), 95% confidence interval 1175.4 to 1200.21,188(1175.4–1200.2)
  103. 103
    OpenAIDossier

    high

    Arena rating

    GPT-5 mini, Arena rating: 1,185 (scale 800.0 to 1309.4), 95% confidence interval 1177.0 to 1193.21,185(1177.0–1193.2)
  104. 104
    AlibabaDossier

    Arena rating

    Qwen VL Max, Arena rating: 1,185 (scale 800.0 to 1309.4), 95% confidence interval 1172.6 to 1196.91,185(1172.6–1196.9)
  105. 105

    Arena rating

    Grok 4 (0709), Arena rating: 1,184 (scale 800.0 to 1309.4), 95% confidence interval 1176.9 to 1191.61,184(1176.9–1191.6)
  106. 106
    GoogleDossier

    Arena rating

    Gemini 1.5 Pro (002), Arena rating: 1,180 (scale 800.0 to 1309.4), 95% confidence interval 1170.5 to 1188.81,180(1170.5–1188.8)
  107. 107
    AnthropicDossier

    Arena rating

    Claude 3.7 Sonnet, Arena rating: 1,175 (scale 800.0 to 1309.4), 95% confidence interval 1166.0 to 1184.71,175(1166.0–1184.7)
  108. 108
    GoogleDossier

    non-thinking

    Arena rating

    Gemini 2.5 Flash-Lite Preview (Sep '25), Arena rating: 1,174 (scale 800.0 to 1309.4), 95% confidence interval 1164.6 to 1183.81,174(1164.6–1183.8)
  109. 109
    GoogleDossier

    Arena rating

    Gemini 2.0 Flash (001), Arena rating: 1,172 (scale 800.0 to 1309.4), 95% confidence interval 1164.9 to 1179.41,172(1164.9–1179.4)
  110. 110
    OpenAIDossier

    Arena rating

    GPT-4o (May '24), Arena rating: 1,162 (scale 800.0 to 1309.4), 95% confidence interval 1153.2 to 1170.81,162(1153.2–1170.8)
  111. 111
    AnthropicDossier

    Arena rating

    Claude 3.5 Sonnet (Oct '24), Arena rating: 1,161 (scale 800.0 to 1309.4), 95% confidence interval 1153.2 to 1169.21,161(1153.2–1169.2)
  112. 112

    Arena rating

    GLM-4.6V, Arena rating: 1,161 (scale 800.0 to 1309.4), 95% confidence interval 1146.8 to 1174.31,161(1146.8–1174.3)
  113. 113
    GoogleDossier

    Arena rating

    Gemma 3 27B, Arena rating: 1,160 (scale 800.0 to 1309.4), 95% confidence interval 1151.7 to 1168.01,160(1151.7–1168.0)
  114. 114
    MistralDossier

    Arena rating

    Mistral Medium 3.1, Arena rating: 1,160 (scale 800.0 to 1309.4), 95% confidence interval 1153.4 to 1165.81,160(1153.4–1165.8)
  115. 115
    StepFunDossier

    Arena rating

    Step-1o Turbo (202506), Arena rating: 1,155 (scale 800.0 to 1309.4), 95% confidence interval 1140.9 to 1169.61,155(1140.9–1169.6)
  116. 116
    MistralDossier

    Arena rating

    Mistral Medium 3, Arena rating: 1,155 (scale 800.0 to 1309.4), 95% confidence interval 1146.8 to 1163.41,155(1146.8–1163.4)
  117. 117
    TencentDossier

    Arena rating

    Hunyuan Vision 1.5 Thinking, Arena rating: 1,154 (scale 800.0 to 1309.4), 95% confidence interval 1141.6 to 1166.21,154(1141.6–1166.2)
  118. 118

    Arena rating

    GLM-4.5V, Arena rating: 1,153 (scale 800.0 to 1309.4), 95% confidence interval 1141.0 to 1164.51,153(1141.0–1164.5)
  119. 119
    TencentDossier

    Arena rating

    Hunyuan Large Vision, Arena rating: 1,150 (scale 800.0 to 1309.4), 95% confidence interval 1133.1 to 1166.21,150(1133.1–1166.2)
  120. 120
    OpenAIDossier

    high

    Arena rating

    GPT-5 nano, Arena rating: 1,147 (scale 800.0 to 1309.4), 95% confidence interval 1136.1 to 1158.81,147(1136.1–1158.8)
  121. 121
    AnthropicDossier

    Arena rating

    Claude 3.5 Sonnet (June '24), Arena rating: 1,146 (scale 800.0 to 1309.4), 95% confidence interval 1135.5 to 1155.71,146(1135.5–1155.7)
  122. 122

    Arena rating

    Llama 4 Maverick, Arena rating: 1,145 (scale 800.0 to 1309.4), 95% confidence interval 1136.3 to 1154.51,145(1136.3–1154.5)
  123. 123
    StepFunDossier

    Arena rating

    Step-3, Arena rating: 1,144 (scale 800.0 to 1309.4), 95% confidence interval 1132.2 to 1156.11,144(1132.2–1156.1)
  124. 124
    GoogleDossier

    Arena rating

    Gemini 1.5 Flash (002), Arena rating: 1,142 (scale 800.0 to 1309.4), 95% confidence interval 1131.8 to 1151.71,142(1131.8–1151.7)
  125. 125
    MistralDossier

    Arena rating

    Mistral Small 3.2, Arena rating: 1,140 (scale 800.0 to 1309.4), 95% confidence interval 1130.7 to 1148.91,140(1130.7–1148.9)
  126. 126
    GoogleDossier

    Arena rating

    Gemini 2.0 Flash-Lite Preview (02-05), Arena rating: 1,136 (scale 800.0 to 1309.4), 95% confidence interval 1125.5 to 1146.51,136(1125.5–1146.5)
  127. 127

    Arena rating

    Llama 4 Scout, Arena rating: 1,128 (scale 800.0 to 1309.4), 95% confidence interval 1118.1 to 1137.21,128(1118.1–1137.2)
  128. 128
    AnthropicDossier

    Arena rating

    Claude 3.5 Haiku, Arena rating: 1,127 (scale 800.0 to 1309.4), 95% confidence interval 1111.2 to 1143.31,127(1111.2–1143.3)
  129. 129
    MistralDossier

    Arena rating

    Mistral Small 3.1, Arena rating: 1,127 (scale 800.0 to 1309.4), 95% confidence interval 1118.4 to 1135.41,127(1118.4–1135.4)
  130. 130
    StepFunDossier

    high-res

    Arena rating

    Step-1o Vision 32K, Arena rating: 1,124 (scale 800.0 to 1309.4), 95% confidence interval 1111.4 to 1136.71,124(1111.4–1136.7)
  131. 131
    AlibabaDossier

    Arena rating

    Qwen2.5 VL 72B Instruct, Arena rating: 1,121 (scale 800.0 to 1309.4), 95% confidence interval 1110.4 to 1132.11,121(1110.4–1132.1)
  132. 132
    GoogleDossier

    Arena rating

    Gemini 1.5 Pro (001), Arena rating: 1,120 (scale 800.0 to 1309.4), 95% confidence interval 1108.1 to 1131.41,120(1108.1–1131.4)
  133. 133
    AlibabaDossier

    Arena rating

    Qwen2.5 VL 32B Instruct, Arena rating: 1,118 (scale 800.0 to 1309.4), 95% confidence interval 1102.8 to 1133.81,118(1102.8–1133.8)
  134. 134
    OpenAIDossier

    Arena rating

    GPT-4o (Aug '24), Arena rating: 1,118 (scale 800.0 to 1309.4), 95% confidence interval 1106.0 to 1130.61,118(1106.0–1130.6)
  135. 135
    Allen Institute for AIDossier

    Arena rating

    Molmo 2 8B, Arena rating: 1,113 (scale 800.0 to 1309.4), 95% confidence interval 1093.4 to 1132.81,113(1093.4–1132.8)
  136. 136
    OpenAIDossier

    Arena rating

    GPT-4 Turbo, Arena rating: 1,112 (scale 800.0 to 1309.4), 95% confidence interval 1100.6 to 1123.81,112(1100.6–1123.8)
  137. 137
    OpenAIDossier

    Arena rating

    GPT-4o mini, Arena rating: 1,097 (scale 800.0 to 1309.4), 95% confidence interval 1088.7 to 1105.41,097(1088.7–1105.4)
  138. 138
    MistralDossier

    Arena rating

    Pixtral Large (2411), Arena rating: 1,095 (scale 800.0 to 1309.4), 95% confidence interval 1086.1 to 1104.81,095(1086.1–1104.8)
  139. 139
    OpenAIDossier

    Arena rating

    GPT-4.1 nano, Arena rating: 1,089 (scale 800.0 to 1309.4), 95% confidence interval 1070.4 to 1107.01,089(1070.4–1107.0)
  140. 140
    AlibabaDossier

    Arena rating

    Qwen2 VL 72B, Arena rating: 1,085 (scale 800.0 to 1309.4), 95% confidence interval 1074.7 to 1095.11,085(1074.7–1095.1)
  141. 141
    AlibabaDossier

    Arena rating

    Qwen VL Max (1119), Arena rating: 1,085 (scale 800.0 to 1309.4), 95% confidence interval 1068.7 to 1100.81,085(1068.7–1100.8)
  142. 142
    GoogleDossier

    Arena rating

    Gemini 1.5 Flash-8B (001), Arena rating: 1,071 (scale 800.0 to 1309.4), 95% confidence interval 1060.8 to 1082.11,071(1060.8–1082.1)
  143. 143
    StepFunDossier

    Arena rating

    Step-1V 32K, Arena rating: 1,064 (scale 800.0 to 1309.4), 95% confidence interval 1047.5 to 1081.11,064(1047.5–1081.1)
  144. 144
    AnthropicDossier

    Arena rating

    Claude 3 Opus, Arena rating: 1,062 (scale 800.0 to 1309.4), 95% confidence interval 1051.5 to 1072.31,062(1051.5–1072.3)
  145. 145
    GoogleDossier

    Arena rating

    Gemini 1.5 Flash (001), Arena rating: 1,059 (scale 800.0 to 1309.4), 95% confidence interval 1047.6 to 1071.41,059(1047.6–1071.4)
  146. 146
    Allen Institute for AIDossier

    Arena rating

    Molmo 72B (0924), Arena rating: 1,045 (scale 800.0 to 1309.4), 95% confidence interval 1032.1 to 1058.31,045(1032.1–1058.3)
  147. 147
    TencentDossier

    Arena rating

    Hunyuan Standard Vision, Arena rating: 1,042 (scale 800.0 to 1309.4), 95% confidence interval 1020.8 to 1062.91,042(1020.8–1062.9)
  148. 148

    Arena rating

    Llama 3.2 90B Vision Instruct, Arena rating: 1,032 (scale 800.0 to 1309.4), 95% confidence interval 1023.2 to 1041.81,032(1023.2–1041.8)
  149. 149
    AlibabaDossier

    Arena rating

    Qwen2 VL 7B Instruct, Arena rating: 1,031 (scale 800.0 to 1309.4), 95% confidence interval 1020.4 to 1041.61,031(1020.4–1041.6)
  150. 150
    MistralDossier

    Arena rating

    Pixtral 12B (2409), Arena rating: 1,025 (scale 800.0 to 1309.4), 95% confidence interval 1015.7 to 1034.91,025(1015.7–1034.9)
  151. 151
    Shanghai AI LaboratoryDossier

    Arena rating

    InternVL2 26B, Arena rating: 1,025 (scale 800.0 to 1309.4), 95% confidence interval 1012.2 to 1037.91,025(1012.2–1037.9)
  152. 152
    AmazonDossier

    Arena rating

    Nova Lite, Arena rating: 1,019 (scale 800.0 to 1309.4), 95% confidence interval 1004.0 to 1034.11,019(1004.0–1034.1)
  153. 153
    AmazonDossier

    Arena rating

    Nova Pro, Arena rating: 1,019 (scale 800.0 to 1309.4), 95% confidence interval 1004.9 to 1032.61,019(1004.9–1032.6)
  154. 154
    AnthropicDossier

    Arena rating

    Claude 3 Sonnet, Arena rating: 1,017 (scale 800.0 to 1309.4), 95% confidence interval 1005.3 to 1028.11,017(1005.3–1028.1)
  155. 155
    01.AIDossier

    Arena rating

    Yi Vision, Arena rating: 1,003 (scale 800.0 to 1309.4), 95% confidence interval 984.6 to 1020.81,003(984.6–1020.8)
  156. 156
    CohereDossier

    Arena rating

    Aya Vision 32B, Arena rating: 1,001 (scale 800.0 to 1309.4), 95% confidence interval 978.5 to 1022.51,001(978.5–1022.5)
  157. 157
    AnthropicDossier

    Arena rating

    Claude 3 Haiku, Arena rating: 1,000 (scale 800.0 to 1309.4), 95% confidence interval 987.7 to 1012.61,000(987.7–1012.6)
  158. 158
    Allen Institute for AIDossier

    Arena rating

    Molmo 7B-D (0924), Arena rating: 994 (scale 800.0 to 1309.4), 95% confidence interval 980.3 to 1007.9994(980.3–1007.9)
  159. 159

    Arena rating

    Llama 3.2 Instruct 11B (Vision), Arena rating: 993 (scale 800.0 to 1309.4), 95% confidence interval 981.6 to 1004.3993(981.6–1004.3)
  160. 160
    NVIDIADossier

    Arena rating

    NVILA Internal 15B v1, Arena rating: 986 (scale 800.0 to 1309.4), 95% confidence interval 965.4 to 1005.7986(965.4–1005.7)
  161. 161
    LLaVA TeamDossier

    Arena rating

    LLaVA-OneVision Qwen2 72B OV, Arena rating: 978 (scale 800.0 to 1309.4), 95% confidence interval 960.2 to 996.6978(960.2–996.6)
  162. 162
    LLaVA TeamDossier

    Arena rating

    LLaVA 1.6 34B, Arena rating: 965 (scale 800.0 to 1309.4), 95% confidence interval 952.4 to 976.9965(952.4–976.9)
  163. 163

    Arena rating

    CogVLM2 Llama3 Chat 19B, Arena rating: 964 (scale 800.0 to 1309.4), 95% confidence interval 948.8 to 979.2964(948.8–979.2)
  164. 164
    OpenBMBDossier

    Arena rating

    MiniCPM-V 2.6, Arena rating: 963 (scale 800.0 to 1309.4), 95% confidence interval 946.9 to 978.6963(946.9–978.6)
  165. 165
    Shanghai AI LaboratoryDossier

    Arena rating

    InternVL2 4B, Arena rating: 959 (scale 800.0 to 1309.4), 95% confidence interval 945.9 to 971.2959(945.9–971.2)
  166. 166
    MicrosoftDossier

    Arena rating

    Phi-3.5 Vision Instruct, Arena rating: 919 (scale 800.0 to 1309.4), 95% confidence interval 903.7 to 935.3919(903.7–935.3)
  167. 167
    MicrosoftDossier

    Arena rating

    Phi-3 Vision 128K Instruct, Arena rating: 881 (scale 800.0 to 1309.4), 95% confidence interval 862.4 to 899.7881(862.4–899.7)