Acropolis

LMArena Text

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

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

No filters. Showing 421 of 421 rows.

  1. 1
    GoogleDossier

    high

    Arena rating

    Gemini 4 Carbon, Arena rating: 1,538 (scale 750.0 to 1538.4), 95% confidence interval 1528.8 to 1547.91,538(1528.8–1547.9)
  2. 2
    GoogleDossier

    high

    Tied #1

    Arena rating

    Gemini 4 Argon, Arena rating: 1,529 (scale 750.0 to 1538.4), 95% confidence interval 1524.1 to 1534.81,529(1524.1–1534.8)
  3. 3
    AnthropicDossier

    high

    Arena rating

    Opus 5.5, Arena rating: 1,514 (scale 750.0 to 1538.4), 95% confidence interval 1507.7 to 1519.61,514(1507.7–1519.6)
  4. 4
    AnthropicDossier

    max

    Arena rating

    Fable 5.1, Arena rating: 1,512 (scale 750.0 to 1538.4), 95% confidence interval 1506.6 to 1517.51,512(1506.6–1517.5)
  5. 5
    AnthropicDossier

    max

    Arena rating

    Opus 5, Arena rating: 1,507 (scale 750.0 to 1538.4), 95% confidence interval 1501.8 to 1511.21,507(1501.8–1511.2)
  6. 6

    max

    Arena rating

    Muse Wonder 1.0, Arena rating: 1,503 (scale 750.0 to 1538.4), 95% confidence interval 1496.8 to 1509.11,503(1496.8–1509.1)
  7. 7
    AnthropicDossier

    high

    Arena rating

    Opus 5, Arena rating: 1,502 (scale 750.0 to 1538.4), 95% confidence interval 1498.6 to 1505.81,502(1498.6–1505.8)
  8. 8
    GoogleDossier

    high

    Arena rating

    Gemini 4 Hydrogen, Arena rating: 1,502 (scale 750.0 to 1538.4), 95% confidence interval 1493.1 to 1511.11,502(1493.1–1511.1)
  9. 9
    AnthropicDossier

    high

    Arena rating

    Opus 4.6, Arena rating: 1,501 (scale 750.0 to 1538.4), 95% confidence interval 1498.2 to 1504.71,501(1498.2–1504.7)
  10. 10
    GoogleDossier

    high

    Arena rating

    Gemini 3.8 Flash, Arena rating: 1,500 (scale 750.0 to 1538.4), 95% confidence interval 1495.4 to 1503.81,500(1495.4–1503.8)
  11. 11
    AnthropicDossier

    Arena rating

    Opus 4.6, Arena rating: 1,497 (scale 750.0 to 1538.4), 95% confidence interval 1493.6 to 1499.91,497(1493.6–1499.9)
  12. 12
    AnthropicDossier

    high

    Arena rating

    Fable 5, Arena rating: 1,492 (scale 750.0 to 1538.4), 95% confidence interval 1487.5 to 1495.81,492(1487.5–1495.8)
  13. 13

    max

    Arena rating

    Muse Spark 1.3, Arena rating: 1,492 (scale 750.0 to 1538.4), 95% confidence interval 1486.7 to 1496.61,492(1486.7–1496.6)
  14. 14
    XiaomiDossier

    Arena rating

    MiMo-V2.6-Pro, Arena rating: 1,492 (scale 750.0 to 1538.4), 95% confidence interval 1482.2 to 1501.01,492(1482.2–1501.0)
  15. 15
    AnthropicDossier

    high

    Arena rating

    Opus 4.7, Arena rating: 1,490 (scale 750.0 to 1538.4), 95% confidence interval 1486.0 to 1493.51,490(1486.0–1493.5)
  16. 16
    GoogleDossier

    high

    Arena rating

    Gemini 3.7 Flash, Arena rating: 1,489 (scale 750.0 to 1538.4), 95% confidence interval 1484.0 to 1493.41,489(1484.0–1493.4)
  17. 17
    AlibabaDossier

    Arena rating

    Qwen3.8-Max, Arena rating: 1,485 (scale 750.0 to 1538.4), 95% confidence interval 1480.7 to 1489.81,485(1480.7–1489.8)
  18. 18

    xhigh

    Arena rating

    Muse Spark 1.2, Arena rating: 1,483 (scale 750.0 to 1538.4), 95% confidence interval 1477.7 to 1489.21,483(1477.7–1489.2)
  19. 19
    AnthropicDossier

    Arena rating

    Opus 4.7, Arena rating: 1,483 (scale 750.0 to 1538.4), 95% confidence interval 1479.5 to 1487.01,483(1479.5–1487.0)
  20. 20
    GoogleDossier

    high

    Arena rating

    Gemini 3.5 Flash, Arena rating: 1,483 (scale 750.0 to 1538.4), 95% confidence interval 1479.0 to 1486.81,483(1479.0–1486.8)
  21. 21
    GoogleDossier

    Arena rating

    Gemini 3.1 Pro Preview, Arena rating: 1,480 (scale 750.0 to 1538.4), 95% confidence interval 1477.1 to 1483.01,480(1477.1–1483.0)
  22. 22
    GoogleDossier

    Arena rating

    Gemini 3 Pro Preview, Arena rating: 1,480 (scale 750.0 to 1538.4), 95% confidence interval 1475.7 to 1483.31,480(1475.7–1483.3)
  23. 23
    GoogleDossier

    high

    Arena rating

    Gemini 3.6 Flash, Arena rating: 1,479 (scale 750.0 to 1538.4), 95% confidence interval 1475.0 to 1483.81,479(1475.0–1483.8)
  24. 24

    Arena rating

    Muse Spark 1.1, Arena rating: 1,479 (scale 750.0 to 1538.4), 95% confidence interval 1474.7 to 1483.01,479(1474.7–1483.0)
  25. 25
    GoogleDossier

    medium

    Arena rating

    Gemini 3.5 Flash, Arena rating: 1,478 (scale 750.0 to 1538.4), 95% confidence interval 1473.7 to 1482.01,478(1473.7–1482.0)
  26. 26

    max

    Arena rating

    GLM-5.3, Arena rating: 1,475 (scale 750.0 to 1538.4), 95% confidence interval 1470.3 to 1479.61,475(1470.3–1479.6)
  27. 27
    AlibabaDossier

    Arena rating

    Qwen3.7 Max Preview, Arena rating: 1,475 (scale 750.0 to 1538.4), 95% confidence interval 1465.0 to 1484.41,475(1465.0–1484.4)
  28. 28
    Moonshot AIDossier

    max

    Arena rating

    Kimi K3, Arena rating: 1,475 (scale 750.0 to 1538.4), 95% confidence interval 1470.3 to 1478.91,475(1470.3–1478.9)
  29. 29

    Arena rating

    Muse Spark, Arena rating: 1,474 (scale 750.0 to 1538.4), 95% confidence interval 1468.5 to 1480.01,474(1468.5–1480.0)
  30. 30

    Arena rating

    GLM-5.3 Flash, Arena rating: 1,472 (scale 750.0 to 1538.4), 95% confidence interval 1467.3 to 1476.11,472(1467.3–1476.1)
  31. 31
    OpenAIDossier

    high

    Arena rating

    GPT-5.5, Arena rating: 1,471 (scale 750.0 to 1538.4), 95% confidence interval 1467.4 to 1474.81,471(1467.4–1474.8)
  32. 32
    AlibabaDossier

    Arena rating

    Qwen3.5 Max Preview, Arena rating: 1,471 (scale 750.0 to 1538.4), 95% confidence interval 1466.0 to 1475.71,471(1466.0–1475.7)
  33. 33

    max

    Arena rating

    GLM-5.2, Arena rating: 1,469 (scale 750.0 to 1538.4), 95% confidence interval 1465.3 to 1473.41,469(1465.3–1473.4)
  34. 34
    AnthropicDossier

    xhigh

    Arena rating

    Sonnet 5.5, Arena rating: 1,469 (scale 750.0 to 1538.4), 95% confidence interval 1463.1 to 1475.11,469(1463.1–1475.1)
  35. 35
    OpenAIDossier

    high

    Arena rating

    GPT-5.4, Arena rating: 1,469 (scale 750.0 to 1538.4), 95% confidence interval 1465.3 to 1472.81,469(1465.3–1472.8)
  36. 36
    BaiduDossier

    Arena rating

    ERNIE 5.1, Arena rating: 1,468 (scale 750.0 to 1538.4), 95% confidence interval 1463.3 to 1472.31,468(1463.3–1472.3)
  37. 37
    OpenAIDossier

    Arena rating

    GPT-5.5, Arena rating: 1,466 (scale 750.0 to 1538.4), 95% confidence interval 1462.5 to 1469.91,466(1462.5–1469.9)
  38. 38
    GoogleDossier

    Arena rating

    Gemini 3 Flash Preview, Arena rating: 1,466 (scale 750.0 to 1538.4), 95% confidence interval 1461.7 to 1470.31,466(1461.7–1470.3)
  39. 39
    XiaomiDossier

    Arena rating

    MiMo-V2.5-Pro, Arena rating: 1,464 (scale 750.0 to 1538.4), 95% confidence interval 1460.4 to 1467.51,464(1460.4–1467.5)
  40. 40
    DeepSeekDossier

    max

    Arena rating

    DeepSeek V4.1 Flash, Arena rating: 1,464 (scale 750.0 to 1538.4), 95% confidence interval 1457.4 to 1469.71,464(1457.4–1469.7)
  41. 41

    Arena rating

    GLM-5.1, Arena rating: 1,462 (scale 750.0 to 1538.4), 95% confidence interval 1458.9 to 1466.01,462(1458.9–1466.0)
  42. 42
    AnthropicDossier

    high

    Arena rating

    Opus 4.8, Arena rating: 1,462 (scale 750.0 to 1538.4), 95% confidence interval 1457.6 to 1465.41,462(1457.6–1465.4)
  43. 43
    Moonshot AIDossier

    max

    Arena rating

    Kimi K3.1 Pro, Arena rating: 1,459 (scale 750.0 to 1538.4), 95% confidence interval 1449.2 to 1469.11,459(1449.2–1469.1)
  44. 44
    AnthropicDossier

    Arena rating

    Sonnet 4.6, Arena rating: 1,458 (scale 750.0 to 1538.4), 95% confidence interval 1454.4 to 1461.41,458(1454.4–1461.4)
  45. 45
    GoogleDossier

    Arena rating

    Gemini 2.5 Pro, Arena rating: 1,458 (scale 750.0 to 1538.4), 95% confidence interval 1455.4 to 1460.31,458(1455.4–1460.3)
  46. 46
    OpenAIDossier

    xhigh

    Arena rating

    GPT-5.6 Sol, Arena rating: 1,457 (scale 750.0 to 1538.4), 95% confidence interval 1453.1 to 1461.61,457(1453.1–1461.6)
  47. 47
    Moonshot AIDossier

    Arena rating

    Kimi K2.6, Arena rating: 1,455 (scale 750.0 to 1538.4), 95% confidence interval 1451.0 to 1459.81,455(1451.0–1459.8)
  48. 48
    AlibabaDossier

    Arena rating

    Qwen3.7 Plus, Arena rating: 1,454 (scale 750.0 to 1538.4), 95% confidence interval 1449.9 to 1458.61,454(1449.9–1458.6)
  49. 49
    AnthropicDossier

    Arena rating

    Opus 4.8, Arena rating: 1,454 (scale 750.0 to 1538.4), 95% confidence interval 1450.0 to 1457.81,454(1450.0–1457.8)
  50. 50
    OpenAIDossier

    Arena rating

    GPT-5.4, Arena rating: 1,453 (scale 750.0 to 1538.4), 95% confidence interval 1448.8 to 1456.21,453(1448.8–1456.2)
  51. 51
    DeepSeekDossier

    Arena rating

    DeepSeek V4-Pro, Arena rating: 1,451 (scale 750.0 to 1538.4), 95% confidence interval 1447.4 to 1455.21,451(1447.4–1455.2)
  52. 52

    thinking

    Arena rating

    Grok 4.20 0309, Arena rating: 1,451 (scale 750.0 to 1538.4), 95% confidence interval 1447.5 to 1454.81,451(1447.5–1454.8)
  53. 53

    Arena rating

    Grok 4.20 Multi-Agent Beta (0309), Arena rating: 1,451 (scale 750.0 to 1538.4), 95% confidence interval 1447.4 to 1454.81,451(1447.4–1454.8)
  54. 54
    AnthropicDossier

    Arena rating

    Opus 4.5, Arena rating: 1,451 (scale 750.0 to 1538.4), 95% confidence interval 1447.9 to 1454.11,451(1447.9–1454.1)
  55. 55
    XiaomiDossier

    Arena rating

    MiMo-V2.6-Flash, Arena rating: 1,451 (scale 750.0 to 1538.4), 95% confidence interval 1445.7 to 1456.01,451(1445.7–1456.0)
  56. 56
    OpenAIDossier

    xhigh

    Arena rating

    GPT-5.6 Terra, Arena rating: 1,449 (scale 750.0 to 1538.4), 95% confidence interval 1444.7 to 1453.01,449(1444.7–1453.0)
  57. 57
    AmazonDossier

    Arena rating

    Nova Experimental Chat (26-02-10), Arena rating: 1,449 (scale 750.0 to 1538.4), 95% confidence interval 1439.0 to 1458.31,449(1439.0–1458.3)
  58. 58
    ByteDanceDossier

    Arena rating

    Dola Seed 2.0 Pro, Arena rating: 1,448 (scale 750.0 to 1538.4), 95% confidence interval 1445.1 to 1451.81,448(1445.1–1451.8)
  59. 59
    AnthropicDossier

    high

    Arena rating

    Opus 4.5, Arena rating: 1,448 (scale 750.0 to 1538.4), 95% confidence interval 1444.5 to 1452.21,448(1444.5–1452.2)
  60. 60

    Arena rating

    Grok 4.5, Arena rating: 1,448 (scale 750.0 to 1538.4), 95% confidence interval 1443.7 to 1452.41,448(1443.7–1452.4)
  61. 61
    AlibabaDossier

    Arena rating

    Qwen3.6 Max Preview, Arena rating: 1,446 (scale 750.0 to 1538.4), 95% confidence interval 1438.1 to 1454.51,446(1438.1–1454.5)
  62. 62

    Arena rating

    GLM-5, Arena rating: 1,446 (scale 750.0 to 1538.4), 95% confidence interval 1441.8 to 1450.31,446(1441.8–1450.3)
  63. 63
    NVIDIADossier

    Arena rating

    Nemotron 3 Ultra 550B A55B (NVFP4), Arena rating: 1,445 (scale 750.0 to 1538.4), 95% confidence interval 1438.2 to 1451.91,445(1438.2–1451.9)
  64. 64
    DeepSeekDossier

    high

    Arena rating

    DeepSeek V4 Pro Preview, Arena rating: 1,445 (scale 750.0 to 1538.4), 95% confidence interval 1441.0 to 1449.01,445(1441.0–1449.0)
  65. 65
    DeepSeekDossier

    high

    Arena rating

    DeepSeek V4 Pro 0813, Arena rating: 1,445 (scale 750.0 to 1538.4), 95% confidence interval 1438.2 to 1451.51,445(1438.2–1451.5)
  66. 66
    OpenAIDossier

    max

    Arena rating

    GPT-6.1 Sol, Arena rating: 1,445 (scale 750.0 to 1538.4), 95% confidence interval 1438.6 to 1450.91,445(1438.6–1450.9)
  67. 67
    Moonshot AIDossier

    thinking

    Arena rating

    Kimi K2.5, Arena rating: 1,445 (scale 750.0 to 1538.4), 95% confidence interval 1441.4 to 1447.91,445(1441.4–1447.9)
  68. 68

    Arena rating

    Grok 4.20 Beta 1, Arena rating: 1,444 (scale 750.0 to 1538.4), 95% confidence interval 1439.9 to 1449.01,444(1439.9–1449.0)
  69. 69
    BaiduDossier

    Arena rating

    ERNIE 5.0 (0110), Arena rating: 1,444 (scale 750.0 to 1538.4), 95% confidence interval 1440.4 to 1448.11,444(1440.4–1448.1)
  70. 70
    StepFunDossier

    Arena rating

    Step 5 Preview, Arena rating: 1,443 (scale 750.0 to 1538.4), 95% confidence interval 1437.1 to 1449.61,443(1437.1–1449.6)
  71. 71
    Thinking MachinesDossier

    Arena rating

    Inkling, Arena rating: 1,443 (scale 750.0 to 1538.4), 95% confidence interval 1439.0 to 1447.11,443(1439.0–1447.1)
  72. 72
    BaiduDossier

    Arena rating

    ERNIE 5.0 Preview (1203), Arena rating: 1,443 (scale 750.0 to 1538.4), 95% confidence interval 1436.1 to 1449.01,443(1436.1–1449.0)
  73. 73
    GoogleDossier

    Arena rating

    Gemma 4 31B, Arena rating: 1,443 (scale 750.0 to 1538.4), 95% confidence interval 1435.1 to 1449.91,443(1435.1–1449.9)
  74. 74
    OpenAIDossier

    high

    Arena rating

    GPT-5.1, Arena rating: 1,443 (scale 750.0 to 1538.4), 95% confidence interval 1438.8 to 1446.21,443(1438.8–1446.2)
  75. 75
    GoogleDossier

    minimal

    Arena rating

    Gemini 3 Flash Preview, Arena rating: 1,442 (scale 750.0 to 1538.4), 95% confidence interval 1439.3 to 1445.41,442(1439.3–1445.4)
  76. 76
    AnthropicDossier

    high

    Arena rating

    Sonnet 5, Arena rating: 1,442 (scale 750.0 to 1538.4), 95% confidence interval 1437.8 to 1445.91,442(1437.8–1445.9)
  77. 77

    Arena rating

    GLM-4.6, Arena rating: 1,440 (scale 750.0 to 1538.4), 95% confidence interval 1436.4 to 1444.21,440(1436.4–1444.2)
  78. 78
    OpenAIDossier

    max

    Arena rating

    GPT-6 Astra, Arena rating: 1,440 (scale 750.0 to 1538.4), 95% confidence interval 1434.1 to 1445.11,440(1434.1–1445.1)
  79. 79
    AlibabaDossier

    Arena rating

    Qwen3 Max Preview, Arena rating: 1,439 (scale 750.0 to 1538.4), 95% confidence interval 1434.6 to 1443.61,439(1434.6–1443.6)
  80. 80
    TencentDossier

    Arena rating

    Hy3, Arena rating: 1,439 (scale 750.0 to 1538.4), 95% confidence interval 1433.0 to 1444.31,439(1433.0–1444.3)
  81. 81
    AnthropicDossier

    Arena rating

    Sonnet 4.5, Arena rating: 1,439 (scale 750.0 to 1538.4), 95% confidence interval 1435.8 to 1441.51,439(1435.8–1441.5)
  82. 82
    OpenAIDossier

    Arena rating

    GPT-5.2 Chat, Arena rating: 1,438 (scale 750.0 to 1538.4), 95% confidence interval 1434.5 to 1442.51,438(1434.5–1442.5)
  83. 83
    AlibabaDossier

    Arena rating

    Qwen3.5 397B A17B, Arena rating: 1,438 (scale 750.0 to 1538.4), 95% confidence interval 1434.5 to 1440.51,438(1434.5–1440.5)
  84. 84

    thinking

    Arena rating

    Grok 4.1, Arena rating: 1,437 (scale 750.0 to 1538.4), 95% confidence interval 1433.6 to 1439.81,437(1433.6–1439.8)
  85. 85
    AlibabaDossier

    Arena rating

    Qwen3.8-27B, Arena rating: 1,437 (scale 750.0 to 1538.4), 95% confidence interval 1432.0 to 1441.11,437(1432.0–1441.1)
  86. 86
    XiaomiDossier

    Arena rating

    MiMo-V2-Pro, Arena rating: 1,436 (scale 750.0 to 1538.4), 95% confidence interval 1431.8 to 1441.11,436(1431.8–1441.1)
  87. 87

    Arena rating

    Grok 4.1, Arena rating: 1,436 (scale 750.0 to 1538.4), 95% confidence interval 1433.1 to 1439.41,436(1433.1–1439.4)
  88. 88
    AlibabaDossier

    Arena rating

    Qwen3.6 Plus, Arena rating: 1,436 (scale 750.0 to 1538.4), 95% confidence interval 1432.0 to 1440.21,436(1432.0–1440.2)
  89. 89

    Arena rating

    GLM-4.7, Arena rating: 1,435 (scale 750.0 to 1538.4), 95% confidence interval 1429.4 to 1441.51,435(1429.4–1441.5)
  90. 90
    GoogleDossier

    Arena rating

    Gemini 3.5 Flash-Lite, Arena rating: 1,435 (scale 750.0 to 1538.4), 95% confidence interval 1430.2 to 1439.21,435(1430.2–1439.2)
  91. 91
    AnthropicDossier

    high

    Arena rating

    Sonnet 4.5, Arena rating: 1,434 (scale 750.0 to 1538.4), 95% confidence interval 1431.6 to 1437.11,434(1431.6–1437.1)
  92. 92
    GoogleDossier

    Arena rating

    Gemma 4 26B A4B, Arena rating: 1,434 (scale 750.0 to 1538.4), 95% confidence interval 1426.6 to 1441.51,434(1426.6–1441.5)
  93. 93
    DeepSeekDossier

    Arena rating

    DeepSeek V4-Flash, Arena rating: 1,432 (scale 750.0 to 1538.4), 95% confidence interval 1428.0 to 1436.01,432(1428.0–1436.0)
  94. 94
    OpenAIDossier

    xhigh

    Arena rating

    GPT-5.6 Luna, Arena rating: 1,431 (scale 750.0 to 1538.4), 95% confidence interval 1426.7 to 1435.41,431(1426.7–1435.4)
  95. 95

    Arena rating

    GLM-4.5, Arena rating: 1,430 (scale 750.0 to 1538.4), 95% confidence interval 1425.3 to 1435.11,430(1425.3–1435.1)
  96. 96

    high

    Arena rating

    Grok 4.6, Arena rating: 1,430 (scale 750.0 to 1538.4), 95% confidence interval 1425.4 to 1434.71,430(1425.4–1434.7)
  97. 97
    BaiduDossier

    Arena rating

    ERNIE 5.0 Preview (1022), Arena rating: 1,430 (scale 750.0 to 1538.4), 95% confidence interval 1420.9 to 1438.41,430(1420.9–1438.4)
  98. 98
    MistralDossier

    Arena rating

    Mistral Large 3, Arena rating: 1,429 (scale 750.0 to 1538.4), 95% confidence interval 1425.9 to 1431.51,429(1425.9–1431.5)
  99. 99
    OpenAIDossier

    Arena rating

    ChatGPT-4o Latest, Arena rating: 1,429 (scale 750.0 to 1538.4), 95% confidence interval 1425.7 to 1431.31,429(1425.7–1431.3)
  100. 100
    MiniMaxDossier

    Arena rating

    MiniMax M3, Arena rating: 1,428 (scale 750.0 to 1538.4), 95% confidence interval 1424.4 to 1431.71,428(1424.4–1431.7)
  101. 101
    DeepSeekDossier

    Arena rating

    DeepSeek R1 0528, Arena rating: 1,428 (scale 750.0 to 1538.4), 95% confidence interval 1422.2 to 1433.51,428(1422.2–1433.5)
  102. 102
    XiaomiDossier

    Arena rating

    MiMo-V2.5, Arena rating: 1,427 (scale 750.0 to 1538.4), 95% confidence interval 1423.2 to 1431.71,427(1423.2–1431.7)
  103. 103
    MeituanDossier

    Arena rating

    LongCat Flash Chat Exp (2602), Arena rating: 1,426 (scale 750.0 to 1538.4), 95% confidence interval 1421.8 to 1430.81,426(1421.8–1430.8)
  104. 104

    Arena rating

    Grok 3 Preview (02-24), Arena rating: 1,426 (scale 750.0 to 1538.4), 95% confidence interval 1421.3 to 1430.01,426(1421.3–1430.0)
  105. 105
    DeepSeekDossier

    thinking

    Arena rating

    DeepSeek V3.2 Exp, Arena rating: 1,425 (scale 750.0 to 1538.4), 95% confidence interval 1418.3 to 1431.51,425(1418.3–1431.5)
  106. 106
    DeepSeekDossier

    high

    Arena rating

    DeepSeek V4 Flash Preview, Arena rating: 1,425 (scale 750.0 to 1538.4), 95% confidence interval 1420.6 to 1428.81,425(1420.6–1428.8)
  107. 107
    MistralDossier

    Arena rating

    Mistral Medium 3.1, Arena rating: 1,425 (scale 750.0 to 1538.4), 95% confidence interval 1422.0 to 1427.11,425(1422.0–1427.1)
  108. 108
    DeepSeekDossier

    Arena rating

    DeepSeek V3.2, Arena rating: 1,425 (scale 750.0 to 1538.4), 95% confidence interval 1421.1 to 1428.01,425(1421.1–1428.0)
  109. 109
    ReflectionDossier

    Arena rating

    Beam 1.0, Arena rating: 1,424 (scale 750.0 to 1538.4), 95% confidence interval 1414.9 to 1434.11,424(1414.9–1434.1)
  110. 110
    XiaomiDossier

    Arena rating

    MiMo-V2-Omni, Arena rating: 1,423 (scale 750.0 to 1538.4), 95% confidence interval 1417.2 to 1428.31,423(1417.2–1428.3)
  111. 111
    MeituanDossier

    Arena rating

    LongCat Flash Chat, Arena rating: 1,423 (scale 750.0 to 1538.4), 95% confidence interval 1416.2 to 1429.01,423(1416.2–1429.0)
  112. 112
    DeepSeekDossier

    Arena rating

    DeepSeek V3.2 Exp, Arena rating: 1,422 (scale 750.0 to 1538.4), 95% confidence interval 1416.0 to 1428.71,422(1416.0–1428.7)
  113. 113
    OpenAIDossier

    Arena rating

    GPT-5.1, Arena rating: 1,422 (scale 750.0 to 1538.4), 95% confidence interval 1418.7 to 1425.81,422(1418.7–1425.8)
  114. 114
    MistralDossier

    Arena rating

    Mistral Medium 3.5, Arena rating: 1,421 (scale 750.0 to 1538.4), 95% confidence interval 1414.4 to 1427.21,421(1414.4–1427.2)
  115. 115
    Moonshot AIDossier

    instant

    Arena rating

    Kimi K2.5, Arena rating: 1,420 (scale 750.0 to 1538.4), 95% confidence interval 1414.0 to 1426.91,420(1414.0–1426.9)
  116. 116
    DeepSeekDossier

    thinking

    Arena rating

    DeepSeek V3.2, Arena rating: 1,420 (scale 750.0 to 1538.4), 95% confidence interval 1416.6 to 1423.71,420(1416.6–1423.7)
  117. 117
    AlibabaDossier

    Arena rating

    Qwen3 VL 235B A22B Instruct, Arena rating: 1,420 (scale 750.0 to 1538.4), 95% confidence interval 1413.5 to 1426.51,420(1413.5–1426.5)
  118. 118
    DeepSeekDossier

    Arena rating

    DeepSeek V3.1, Arena rating: 1,420 (scale 750.0 to 1538.4), 95% confidence interval 1413.9 to 1426.01,420(1413.9–1426.0)
  119. 119
    AnthropicDossier

    thinking · 16,000 budget

    Arena rating

    Opus 4.1, Arena rating: 1,419 (scale 750.0 to 1538.4), 95% confidence interval 1415.9 to 1422.81,419(1415.9–1422.8)
  120. 120
    OpenAIDossier

    Arena rating

    GPT-5.5 Instant, Arena rating: 1,419 (scale 750.0 to 1538.4), 95% confidence interval 1414.3 to 1424.21,419(1414.3–1424.2)
  121. 121
    AmazonDossier

    Arena rating

    Nova Experimental Chat (12-10), Arena rating: 1,419 (scale 750.0 to 1538.4), 95% confidence interval 1409.7 to 1428.61,419(1409.7–1428.6)
  122. 122
    DeepSeekDossier

    thinking

    Arena rating

    DeepSeek V3.1 Terminus, Arena rating: 1,419 (scale 750.0 to 1538.4), 95% confidence interval 1409.1 to 1429.31,419(1409.1–1429.3)
  123. 123
    AlibabaDossier

    Arena rating

    Qwen3 235B A22B 2507 Instruct, Arena rating: 1,419 (scale 750.0 to 1538.4), 95% confidence interval 1416.6 to 1421.71,419(1416.6–1421.7)
  124. 124
    AnthropicDossier

    Arena rating

    Opus 4.1, Arena rating: 1,419 (scale 750.0 to 1538.4), 95% confidence interval 1415.6 to 1421.61,419(1415.6–1421.6)
  125. 125
    AlibabaDossier

    Arena rating

    Qwen3.5 122B A10B, Arena rating: 1,417 (scale 750.0 to 1538.4), 95% confidence interval 1413.2 to 1421.71,417(1413.2–1421.7)
  126. 126
    OpenAIDossier

    Arena rating

    GPT-4.5 Preview, Arena rating: 1,417 (scale 750.0 to 1538.4), 95% confidence interval 1411.7 to 1423.11,417(1411.7–1423.1)
  127. 127
    DeepSeekDossier

    Arena rating

    DeepSeek V3.1 Terminus, Arena rating: 1,417 (scale 750.0 to 1538.4), 95% confidence interval 1407.5 to 1427.01,417(1407.5–1427.0)
  128. 128
    AlibabaDossier

    Arena rating

    Qwen3 Next 80B A3B Instruct, Arena rating: 1,417 (scale 750.0 to 1538.4), 95% confidence interval 1412.4 to 1421.91,417(1412.4–1421.9)
  129. 129
    GoogleDossier

    Arena rating

    Gemini 2.5 Flash, Arena rating: 1,417 (scale 750.0 to 1538.4), 95% confidence interval 1414.6 to 1419.51,417(1414.6–1419.5)
  130. 130
    DeepSeekDossier

    thinking

    Arena rating

    DeepSeek V3.1, Arena rating: 1,417 (scale 750.0 to 1538.4), 95% confidence interval 1410.1 to 1423.31,417(1410.1–1423.3)
  131. 131
    OpenAIDossier

    high

    Arena rating

    GPT-5.2, Arena rating: 1,416 (scale 750.0 to 1538.4), 95% confidence interval 1412.7 to 1419.71,416(1412.7–1419.7)
  132. 132
    GoogleDossier

    Arena rating

    Gemini 3.1 Flash-Lite Preview, Arena rating: 1,416 (scale 750.0 to 1538.4), 95% confidence interval 1412.4 to 1419.61,416(1412.4–1419.6)
  133. 133
    AlibabaDossier

    Arena rating

    Qwen3 235B A22B Thinking (2507), Arena rating: 1,415 (scale 750.0 to 1538.4), 95% confidence interval 1408.9 to 1422.01,415(1408.9–1422.0)
  134. 134
    AmazonDossier

    Arena rating

    Nova Experimental Chat (11-10), Arena rating: 1,415 (scale 750.0 to 1538.4), 95% confidence interval 1410.5 to 1418.91,415(1410.5–1418.9)
  135. 135
    Moonshot AIDossier

    Arena rating

    Kimi K2 Thinking Turbo, Arena rating: 1,414 (scale 750.0 to 1538.4), 95% confidence interval 1411.3 to 1417.41,414(1411.3–1417.4)
  136. 136
    Thinking MachinesDossier

    Arena rating

    Inkling Small, Arena rating: 1,414 (scale 750.0 to 1538.4), 95% confidence interval 1409.7 to 1418.71,414(1409.7–1418.7)
  137. 137
    AlibabaDossier

    Arena rating

    Qwen3 Max, Arena rating: 1,413 (scale 750.0 to 1538.4), 95% confidence interval 1406.3 to 1419.21,413(1406.3–1419.2)
  138. 138
    OpenAIDossier

    Arena rating

    GPT-5.2, Arena rating: 1,412 (scale 750.0 to 1538.4), 95% confidence interval 1408.9 to 1415.21,412(1408.9–1415.2)
  139. 139
    OpenAIDossier

    high

    Arena rating

    GPT-5.4 mini, Arena rating: 1,412 (scale 750.0 to 1538.4), 95% confidence interval 1408.0 to 1415.51,412(1408.0–1415.5)
  140. 140

    Arena rating

    Grok 4 (0709), Arena rating: 1,412 (scale 750.0 to 1538.4), 95% confidence interval 1407.6 to 1415.51,412(1407.6–1415.5)
  141. 141
    MistralDossier

    Arena rating

    Mistral Large 4, Arena rating: 1,411 (scale 750.0 to 1538.4), 95% confidence interval 1402.8 to 1419.81,411(1402.8–1419.8)
  142. 142
    XiaomiDossier

    non-thinking

    Arena rating

    MiMo-V2-Flash, Arena rating: 1,411 (scale 750.0 to 1538.4), 95% confidence interval 1407.8 to 1414.71,411(1407.8–1414.7)
  143. 143
    OpenAIDossier

    Arena rating

    o3, Arena rating: 1,410 (scale 750.0 to 1538.4), 95% confidence interval 1406.2 to 1413.41,410(1406.2–1413.4)
  144. 144
    AlibabaDossier

    Arena rating

    Qwen3.5 27B, Arena rating: 1,409 (scale 750.0 to 1538.4), 95% confidence interval 1404.4 to 1413.01,409(1404.4–1413.0)
  145. 145

    thinking

    Arena rating

    Grok 4.1 Fast, Arena rating: 1,408 (scale 750.0 to 1538.4), 95% confidence interval 1405.0 to 1411.41,408(1405.0–1411.4)
  146. 146
    NVIDIADossier

    Arena rating

    Nemotron 3.5 Super VL 120B A12B (NVFP4), Arena rating: 1,407 (scale 750.0 to 1538.4), 95% confidence interval 1397.5 to 1416.61,407(1397.5–1416.6)
  147. 147

    non-thinking

    Arena rating

    Grok 4 Fast, Arena rating: 1,407 (scale 750.0 to 1538.4), 95% confidence interval 1399.2 to 1414.71,407(1399.2–1414.7)
  148. 148
    GoogleDossier

    Arena rating

    Gemini 2.5 Flash Preview (09-2025), Arena rating: 1,407 (scale 750.0 to 1538.4), 95% confidence interval 1402.6 to 1410.61,407(1402.6–1410.6)
  149. 149
    OpenAIDossier

    high

    Arena rating

    GPT-5, Arena rating: 1,406 (scale 750.0 to 1538.4), 95% confidence interval 1402.0 to 1410.91,406(1402.0–1410.9)
  150. 150
    TencentDossier

    Arena rating

    Hunyuan Vision 1.5 Thinking, Arena rating: 1,406 (scale 750.0 to 1538.4), 95% confidence interval 1393.9 to 1418.81,406(1393.9–1418.8)
  151. 151
    MiniMaxDossier

    Arena rating

    MiniMax M2.7, Arena rating: 1,405 (scale 750.0 to 1538.4), 95% confidence interval 1401.3 to 1407.91,405(1401.3–1407.9)
  152. 152
    OpenAIDossier

    Arena rating

    GPT-5 Chat, Arena rating: 1,404 (scale 750.0 to 1538.4), 95% confidence interval 1399.2 to 1407.81,404(1399.2–1407.8)
  153. 153
    Thinking MachinesDossier

    Arena rating

    Inkling 2 Small, Arena rating: 1,403 (scale 750.0 to 1538.4), 95% confidence interval 1390.9 to 1414.41,403(1390.9–1414.4)
  154. 154
    StepFunDossier

    Arena rating

    Step 3.5 Flash, Arena rating: 1,402 (scale 750.0 to 1538.4), 95% confidence interval 1398.9 to 1405.91,402(1398.9–1405.9)
  155. 155
    TencentDossier

    Arena rating

    Hy3 Preview, Arena rating: 1,401 (scale 750.0 to 1538.4), 95% confidence interval 1393.9 to 1408.71,401(1393.9–1408.7)
  156. 156
    TencentDossier

    Arena rating

    Hunyuan T1, Arena rating: 1,401 (scale 750.0 to 1538.4), 95% confidence interval 1392.1 to 1409.41,401(1392.1–1409.4)
  157. 157
    AlibabaDossier

    Arena rating

    Qwen3 VL 235B A22B Thinking, Arena rating: 1,401 (scale 750.0 to 1538.4), 95% confidence interval 1393.7 to 1407.31,401(1393.7–1407.3)
  158. 158

    xhigh

    Arena rating

    Grok 4.7, Arena rating: 1,400 (scale 750.0 to 1538.4), 95% confidence interval 1394.7 to 1405.21,400(1394.7–1405.2)
  159. 159

    thinking

    Arena rating

    Grok 4 Fast, Arena rating: 1,400 (scale 750.0 to 1538.4), 95% confidence interval 1394.6 to 1404.71,400(1394.6–1404.7)
  160. 160
    AnthropicDossier

    high

    Arena rating

    Haiku 5.5, Arena rating: 1,398 (scale 750.0 to 1538.4), 95% confidence interval 1391.5 to 1403.81,398(1391.5–1403.8)
  161. 161
    AlibabaDossier

    Arena rating

    Qwen3.5 Flash, Arena rating: 1,397 (scale 750.0 to 1538.4), 95% confidence interval 1393.4 to 1400.71,397(1393.4–1400.7)
  162. 162

    Arena rating

    Grok 4.3, Arena rating: 1,397 (scale 750.0 to 1538.4), 95% confidence interval 1393.3 to 1400.61,397(1393.3–1400.6)
  163. 163
    OpenAIDossier

    max

    Arena rating

    GPT-6 Sol, Arena rating: 1,397 (scale 750.0 to 1538.4), 95% confidence interval 1391.3 to 1402.21,397(1391.3–1402.2)
  164. 164
    AnthropicDossier

    Arena rating

    Haiku 4.5, Arena rating: 1,396 (scale 750.0 to 1538.4), 95% confidence interval 1394.1 to 1398.91,396(1394.1–1398.9)
  165. 165
    AmazonDossier

    Arena rating

    Nova Experimental Chat (26-01-10), Arena rating: 1,395 (scale 750.0 to 1538.4), 95% confidence interval 1385.2 to 1404.41,395(1385.2–1404.4)
  166. 166
    AlibabaDossier

    Arena rating

    Qwen3.5 35B A3B, Arena rating: 1,395 (scale 750.0 to 1538.4), 95% confidence interval 1390.5 to 1398.91,395(1390.5–1398.9)
  167. 167
    XiaomiDossier

    thinking

    Arena rating

    MiMo-V2-Flash, Arena rating: 1,395 (scale 750.0 to 1538.4), 95% confidence interval 1388.5 to 1400.81,395(1388.5–1400.8)
  168. 168
    AmazonDossier

    Arena rating

    Nova Experimental Chat (10-20), Arena rating: 1,394 (scale 750.0 to 1538.4), 95% confidence interval 1387.8 to 1399.91,394(1387.8–1399.9)
  169. 169
    AlibabaDossier

    non-thinking

    Arena rating

    Qwen3 235B A22B, Arena rating: 1,394 (scale 750.0 to 1538.4), 95% confidence interval 1389.1 to 1398.11,394(1389.1–1398.1)
  170. 170
    MiniMaxDossier

    Arena rating

    MiniMax M2.1 Preview, Arena rating: 1,392 (scale 750.0 to 1538.4), 95% confidence interval 1386.9 to 1397.11,392(1386.9–1397.1)
  171. 171
    OpenAIDossier

    Arena rating

    GPT-5.3 Chat, Arena rating: 1,389 (scale 750.0 to 1538.4), 95% confidence interval 1384.9 to 1393.31,389(1384.9–1393.3)
  172. 172

    Arena rating

    Muse Glimmer, Arena rating: 1,389 (scale 750.0 to 1538.4), 95% confidence interval 1379.0 to 1398.61,389(1379.0–1398.6)
  173. 173
    OpenAIDossier

    max

    Arena rating

    GPT-6 Luna, Arena rating: 1,388 (scale 750.0 to 1538.4), 95% confidence interval 1384.4 to 1392.61,388(1384.4–1392.6)
  174. 174
    UpstageDossier

    Arena rating

    Solar Pro 4, Arena rating: 1,388 (scale 750.0 to 1538.4), 95% confidence interval 1383.2 to 1393.21,388(1383.2–1393.2)
  175. 175
    AlibabaDossier

    Arena rating

    Qwen3 30B A3B 2507 Instruct, Arena rating: 1,384 (scale 750.0 to 1538.4), 95% confidence interval 1378.9 to 1388.71,384(1378.9–1388.7)
  176. 176

    Arena rating

    GLM-4.5-Air, Arena rating: 1,384 (scale 750.0 to 1538.4), 95% confidence interval 1379.5 to 1388.01,384(1379.5–1388.0)
  177. 177
    OpenAIDossier

    Arena rating

    GPT-4.1, Arena rating: 1,383 (scale 750.0 to 1538.4), 95% confidence interval 1379.2 to 1386.61,383(1379.2–1386.6)
  178. 178
    Moonshot AIDossier

    Arena rating

    Kimi K2 0905, Arena rating: 1,380 (scale 750.0 to 1538.4), 95% confidence interval 1373.3 to 1386.31,380(1373.3–1386.3)
  179. 179
    GoogleDossier

    non-thinking

    Arena rating

    Gemini 2.5 Flash-Lite Preview (Sep '25), Arena rating: 1,379 (scale 750.0 to 1538.4), 95% confidence interval 1375.9 to 1382.71,379(1375.9–1382.7)
  180. 180
    TencentDossier

    Arena rating

    Hunyuan TurboS, Arena rating: 1,377 (scale 750.0 to 1538.4), 95% confidence interval 1370.8 to 1383.71,377(1370.8–1383.7)
  181. 181
    AnthropicDossier

    thinking · 16,000 budget

    Arena rating

    Opus 4, Arena rating: 1,377 (scale 750.0 to 1538.4), 95% confidence interval 1372.7 to 1381.61,377(1372.7–1381.6)
  182. 182
    NVIDIADossier

    Arena rating

    Nemotron 3 Super 120B A12B, Arena rating: 1,377 (scale 750.0 to 1538.4), 95% confidence interval 1370.0 to 1384.21,377(1370.0–1384.2)
  183. 183

    Arena rating

    GLM-4.6V, Arena rating: 1,377 (scale 750.0 to 1538.4), 95% confidence interval 1365.3 to 1387.71,377(1365.3–1387.7)
  184. 184
    DeepSeekDossier

    Arena rating

    DeepSeek V3 0324, Arena rating: 1,375 (scale 750.0 to 1538.4), 95% confidence interval 1370.8 to 1378.51,375(1370.8–1378.5)
  185. 185
    OpenAIDossier

    high

    Arena rating

    GPT-5 mini, Arena rating: 1,373 (scale 750.0 to 1538.4), 95% confidence interval 1368.8 to 1377.91,373(1368.8–1377.9)
  186. 186
    DeepSeekDossier

    Arena rating

    DeepSeek R1 (Jan '25), Arena rating: 1,373 (scale 750.0 to 1538.4), 95% confidence interval 1367.9 to 1377.61,373(1367.9–1377.6)
  187. 187
    OpenAIDossier

    high

    Arena rating

    GPT-5.4 nano, Arena rating: 1,372 (scale 750.0 to 1538.4), 95% confidence interval 1368.0 to 1375.51,372(1368.0–1375.5)
  188. 188
    Moonshot AIDossier

    Arena rating

    Kimi K2, Arena rating: 1,371 (scale 750.0 to 1538.4), 95% confidence interval 1366.1 to 1375.91,371(1366.1–1375.9)
  189. 189
    MistralDossier

    Arena rating

    Mistral Medium 3, Arena rating: 1,370 (scale 750.0 to 1538.4), 95% confidence interval 1365.5 to 1374.91,370(1365.5–1374.9)
  190. 190

    high

    Arena rating

    Grok 3 mini, Arena rating: 1,370 (scale 750.0 to 1538.4), 95% confidence interval 1364.2 to 1375.11,370(1364.2–1375.1)
  191. 191
    GoogleDossier

    thinking

    Arena rating

    Gemini 2.5 Flash-Lite Preview (06-17), Arena rating: 1,369 (scale 750.0 to 1538.4), 95% confidence interval 1364.9 to 1373.91,369(1364.9–1373.9)
  192. 192
    AlibabaDossier

    Arena rating

    Qwen3 Next 80B A3B Thinking, Arena rating: 1,368 (scale 750.0 to 1538.4), 95% confidence interval 1362.5 to 1374.21,368(1362.5–1374.2)
  193. 193
    AlibabaDossier

    Arena rating

    Qwen2.5 Max, Arena rating: 1,367 (scale 750.0 to 1538.4), 95% confidence interval 1362.6 to 1370.81,367(1362.6–1370.8)
  194. 194
    AlibabaDossier

    Arena rating

    Qwen3 235B A22B, Arena rating: 1,366 (scale 750.0 to 1538.4), 95% confidence interval 1361.6 to 1371.01,366(1361.6–1371.0)
  195. 195
    AnthropicDossier

    Arena rating

    Opus 4, Arena rating: 1,366 (scale 750.0 to 1538.4), 95% confidence interval 1361.7 to 1370.41,366(1361.7–1370.4)
  196. 196
    OpenAIDossier

    Arena rating

    o1, Arena rating: 1,366 (scale 750.0 to 1538.4), 95% confidence interval 1361.4 to 1370.31,366(1361.4–1370.3)
  197. 197
    OpenAIDossier

    Arena rating

    gpt-oss-120b, Arena rating: 1,366 (scale 750.0 to 1538.4), 95% confidence interval 1361.2 to 1369.91,366(1361.2–1369.9)
  198. 198

    Arena rating

    Grok 3 Mini Beta, Arena rating: 1,364 (scale 750.0 to 1538.4), 95% confidence interval 1359.3 to 1369.51,364(1359.3–1369.5)
  199. 199
    AmazonDossier

    Arena rating

    Nova Experimental Chat (10-09), Arena rating: 1,364 (scale 750.0 to 1538.4), 95% confidence interval 1353.0 to 1374.81,364(1353.0–1374.8)
  200. 200
    AmazonDossier

    Arena rating

    Nova 2.0 Lite, Arena rating: 1,362 (scale 750.0 to 1538.4), 95% confidence interval 1355.8 to 1367.91,362(1355.8–1367.9)
  201. 201

    Arena rating

    Granite 4.2 30B, Arena rating: 1,362 (scale 750.0 to 1538.4), 95% confidence interval 1351.0 to 1372.11,362(1351.0–1372.1)
  202. 202
    InclusionAIDossier

    Arena rating

    Ling Flash 2.0, Arena rating: 1,361 (scale 750.0 to 1538.4), 95% confidence interval 1353.7 to 1368.21,361(1353.7–1368.2)
  203. 203
    MiniMaxDossier

    Arena rating

    MiniMax M2.5, Arena rating: 1,359 (scale 750.0 to 1538.4), 95% confidence interval 1355.4 to 1363.11,359(1355.4–1363.1)
  204. 204
    GoogleDossier

    Arena rating

    Gemma 3 27B, Arena rating: 1,358 (scale 750.0 to 1538.4), 95% confidence interval 1354.1 to 1361.51,358(1354.1–1361.5)
  205. 205
    AlibabaDossier

    Arena rating

    Qwen3 Coder 480B A35B Instruct, Arena rating: 1,357 (scale 750.0 to 1538.4), 95% confidence interval 1351.5 to 1361.51,357(1351.5–1361.5)
  206. 206
    Prime IntellectDossier

    Arena rating

    INTELLECT-3, Arena rating: 1,356 (scale 750.0 to 1538.4), 95% confidence interval 1347.8 to 1364.31,356(1347.8–1364.3)
  207. 207
    InceptionDossier

    Arena rating

    Mercury 2, Arena rating: 1,355 (scale 750.0 to 1538.4), 95% confidence interval 1344.7 to 1365.51,355(1344.7–1365.5)
  208. 208
    GoogleDossier

    Arena rating

    Gemini 2.0 Flash (001), Arena rating: 1,354 (scale 750.0 to 1538.4), 95% confidence interval 1350.5 to 1358.01,354(1350.5–1358.0)
  209. 209
    OpenAIDossier

    Arena rating

    o4-mini, Arena rating: 1,353 (scale 750.0 to 1538.4), 95% confidence interval 1349.2 to 1357.11,353(1349.2–1357.1)
  210. 210
    OpenAIDossier

    Arena rating

    o1-preview, Arena rating: 1,353 (scale 750.0 to 1538.4), 95% confidence interval 1347.8 to 1357.81,353(1347.8–1357.8)
  211. 211
    AnthropicDossier

    thinking · 32,000 budget

    Arena rating

    Claude Sonnet 4, Arena rating: 1,351 (scale 750.0 to 1538.4), 95% confidence interval 1346.4 to 1355.41,351(1346.4–1355.4)
  212. 212

    Arena rating

    GLM-4.7-Flash, Arena rating: 1,351 (scale 750.0 to 1538.4), 95% confidence interval 1344.9 to 1356.31,351(1344.9–1356.3)
  213. 213
    StepFunDossier

    Arena rating

    Step-3, Arena rating: 1,350 (scale 750.0 to 1538.4), 95% confidence interval 1342.6 to 1357.51,350(1342.6–1357.5)
  214. 214
    NVIDIADossier

    Arena rating

    Nemotron 3 Nano 30B A3B (BF16), Arena rating: 1,348 (scale 750.0 to 1538.4), 95% confidence interval 1342.6 to 1353.41,348(1342.6–1353.4)
  215. 215
    MiniMaxDossier

    Arena rating

    MiniMax M1, Arena rating: 1,343 (scale 750.0 to 1538.4), 95% confidence interval 1338.3 to 1346.91,343(1338.3–1346.9)
  216. 216
    OpenAIDossier

    Arena rating

    GPT-4.1 mini, Arena rating: 1,341 (scale 750.0 to 1538.4), 95% confidence interval 1336.3 to 1344.91,341(1336.3–1344.9)
  217. 217
    MiniMaxDossier

    Arena rating

    MiniMax M2, Arena rating: 1,340 (scale 750.0 to 1538.4), 95% confidence interval 1332.7 to 1348.11,340(1332.7–1348.1)
  218. 218
    AlibabaDossier

    Arena rating

    Qwen3 32B, Arena rating: 1,340 (scale 750.0 to 1538.4), 95% confidence interval 1330.8 to 1349.41,340(1330.8–1349.4)
  219. 219
    Arcee AIDossier

    Arena rating

    Trinity Large Thinking, Arena rating: 1,340 (scale 750.0 to 1538.4), 95% confidence interval 1335.3 to 1344.61,340(1335.3–1344.6)
  220. 220
    AnthropicDossier

    Arena rating

    Claude Sonnet 4, Arena rating: 1,339 (scale 750.0 to 1538.4), 95% confidence interval 1334.8 to 1343.71,339(1334.8–1343.7)
  221. 221
    Arcee AIDossier

    Arena rating

    Trinity Large Preview, Arena rating: 1,339 (scale 750.0 to 1538.4), 95% confidence interval 1334.5 to 1343.01,339(1334.5–1343.0)
  222. 222
    NVIDIADossier

    Arena rating

    Llama 3.3 Nemotron Super 49B v1.5, Arena rating: 1,338 (scale 750.0 to 1538.4), 95% confidence interval 1328.4 to 1348.41,338(1328.4–1348.4)
  223. 223
    MistralDossier

    Arena rating

    Mistral Small 3.2, Arena rating: 1,338 (scale 750.0 to 1538.4), 95% confidence interval 1333.2 to 1343.61,338(1333.2–1343.6)
  224. 224
    OpenAIDossier

    high

    Arena rating

    o3-mini, Arena rating: 1,337 (scale 750.0 to 1538.4), 95% confidence interval 1331.4 to 1341.91,337(1331.4–1341.9)
  225. 225
    StepFunDossier

    Arena rating

    Step-1o Turbo (202506), Arena rating: 1,336 (scale 750.0 to 1538.4), 95% confidence interval 1328.8 to 1342.51,336(1328.8–1342.5)
  226. 226
    GoogleDossier

    Arena rating

    Gemma 3 12B, Arena rating: 1,334 (scale 750.0 to 1538.4), 95% confidence interval 1324.9 to 1343.61,334(1324.9–1343.6)
  227. 227
    InclusionAIDossier

    Arena rating

    Ring-flash-2.0, Arena rating: 1,334 (scale 750.0 to 1538.4), 95% confidence interval 1326.4 to 1340.81,334(1326.4–1340.8)
  228. 228

    Arena rating

    GLM-4.5V, Arena rating: 1,333 (scale 750.0 to 1538.4), 95% confidence interval 1324.3 to 1341.21,333(1324.3–1341.2)
  229. 229
    DeepSeekDossier

    Arena rating

    DeepSeek V3, Arena rating: 1,333 (scale 750.0 to 1538.4), 95% confidence interval 1327.8 to 1337.31,333(1327.8–1337.3)
  230. 230
    CohereDossier

    Arena rating

    Command A, Arena rating: 1,331 (scale 750.0 to 1538.4), 95% confidence interval 1327.5 to 1334.41,331(1327.5–1334.4)
  231. 231

    Arena rating

    GLM-4-Plus (0111), Arena rating: 1,331 (scale 750.0 to 1538.4), 95% confidence interval 1322.5 to 1339.31,331(1322.5–1339.3)
  232. 232
    GoogleDossier

    Arena rating

    Gemini 2.0 Flash-Lite Preview (02-05), Arena rating: 1,330 (scale 750.0 to 1538.4), 95% confidence interval 1325.3 to 1333.91,330(1325.3–1333.9)
  233. 233
    NVIDIADossier

    Arena rating

    Nemotron 3.5 Lightning 30B A3B (NVFP4), Arena rating: 1,329 (scale 750.0 to 1538.4), 95% confidence interval 1324.2 to 1334.81,329(1324.2–1334.8)
  234. 234
    AlibabaDossier

    Arena rating

    QwQ 32B, Arena rating: 1,329 (scale 750.0 to 1538.4), 95% confidence interval 1324.6 to 1333.51,329(1324.6–1333.5)
  235. 235
    AlibabaDossier

    Arena rating

    Qwen Plus (0125), Arena rating: 1,326 (scale 750.0 to 1538.4), 95% confidence interval 1318.2 to 1334.71,326(1318.2–1334.7)
  236. 236
    StepFunDossier

    Arena rating

    Step-2 16K Exp (202412), Arena rating: 1,321 (scale 750.0 to 1538.4), 95% confidence interval 1312.4 to 1329.51,321(1312.4–1329.5)
  237. 237
    OpenAIDossier

    high

    Arena rating

    GPT-5 nano, Arena rating: 1,320 (scale 750.0 to 1538.4), 95% confidence interval 1313.1 to 1326.81,320(1313.1–1326.8)
  238. 238
    TencentDossier

    Arena rating

    Hunyuan TurboS, Arena rating: 1,320 (scale 750.0 to 1538.4), 95% confidence interval 1308.1 to 1331.51,320(1308.1–1331.5)
  239. 239
    NVIDIADossier

    Arena rating

    Llama 3.1 Nemotron Ultra 253B v1, Arena rating: 1,320 (scale 750.0 to 1538.4), 95% confidence interval 1308.1 to 1331.11,320(1308.1–1331.1)
  240. 240
    OpenAIDossier

    Arena rating

    o3-mini, Arena rating: 1,320 (scale 750.0 to 1538.4), 95% confidence interval 1316.0 to 1323.11,320(1316.0–1323.1)
  241. 241
    GoogleDossier

    Arena rating

    Gemini 1.5 Pro (002), Arena rating: 1,319 (scale 750.0 to 1538.4), 95% confidence interval 1315.8 to 1322.61,319(1315.8–1322.6)
  242. 242

    Arena rating

    Granite 4.2 8B, Arena rating: 1,319 (scale 750.0 to 1538.4), 95% confidence interval 1307.4 to 1329.71,319(1307.4–1329.7)
  243. 243
    OpenAIDossier

    Arena rating

    o1-mini, Arena rating: 1,317 (scale 750.0 to 1538.4), 95% confidence interval 1313.5 to 1320.81,317(1313.5–1320.8)
  244. 244
    AlibabaDossier

    Arena rating

    Qwen3 30B A3B, Arena rating: 1,317 (scale 750.0 to 1538.4), 95% confidence interval 1312.2 to 1321.61,317(1312.2–1321.6)
  245. 245
    AnthropicDossier

    thinking · 32,000 budget

    Arena rating

    Claude 3.7 Sonnet, Arena rating: 1,314 (scale 750.0 to 1538.4), 95% confidence interval 1310.1 to 1318.61,314(1310.1–1318.6)
  246. 246
    TencentDossier

    Arena rating

    Hunyuan Turbo (0110), Arena rating: 1,311 (scale 750.0 to 1538.4), 95% confidence interval 1300.1 to 1322.91,311(1300.1–1322.9)
  247. 247
    Allen Institute for AIDossier

    Arena rating

    Olmo 3.1 32B Instruct, Arena rating: 1,311 (scale 750.0 to 1538.4), 95% confidence interval 1305.2 to 1317.41,311(1305.2–1317.4)
  248. 248
    NVIDIADossier

    Arena rating

    Llama 3.3 Nemotron Super 49B v1, Arena rating: 1,308 (scale 750.0 to 1538.4), 95% confidence interval 1296.4 to 1320.51,308(1296.4–1320.5)
  249. 249
    GoogleDossier

    Arena rating

    Gemma 3n E4B IT, Arena rating: 1,306 (scale 750.0 to 1538.4), 95% confidence interval 1300.5 to 1310.81,306(1300.5–1310.8)
  250. 250

    Arena rating

    Grok 2, Arena rating: 1,305 (scale 750.0 to 1538.4), 95% confidence interval 1301.0 to 1308.31,305(1301.0–1308.3)
  251. 251
    01.AIDossier

    Arena rating

    Yi Lightning, Arena rating: 1,302 (scale 750.0 to 1538.4), 95% confidence interval 1296.7 to 1306.51,302(1296.7–1306.5)
  252. 252
    OpenAIDossier

    Arena rating

    GPT-4o (May '24), Arena rating: 1,300 (scale 750.0 to 1538.4), 95% confidence interval 1297.0 to 1303.91,300(1297.0–1303.9)
  253. 253
    AnthropicDossier

    Arena rating

    Claude 3.7 Sonnet, Arena rating: 1,299 (scale 750.0 to 1538.4), 95% confidence interval 1295.4 to 1303.31,299(1295.4–1303.3)
  254. 254
    AlibabaDossier

    Arena rating

    Qwen2.5 Plus (1127), Arena rating: 1,299 (scale 750.0 to 1538.4), 95% confidence interval 1292.9 to 1305.51,299(1292.9–1305.5)
  255. 255
    Allen Institute for AIDossier

    Arena rating

    Olmo 3 32B Think, Arena rating: 1,299 (scale 750.0 to 1538.4), 95% confidence interval 1290.8 to 1307.31,299(1290.8–1307.3)
  256. 256
    AnthropicDossier

    Arena rating

    Claude 3.5 Sonnet (Oct '24), Arena rating: 1,298 (scale 750.0 to 1538.4), 95% confidence interval 1294.7 to 1301.11,298(1294.7–1301.1)
  257. 257
    DeepSeekDossier

    Arena rating

    DeepSeek V2.5 (1210), Arena rating: 1,294 (scale 750.0 to 1538.4), 95% confidence interval 1285.7 to 1302.21,294(1285.7–1302.2)
  258. 258

    Arena rating

    Granite 4.2 3B, Arena rating: 1,292 (scale 750.0 to 1538.4), 95% confidence interval 1280.8 to 1304.11,292(1280.8–1304.1)
  259. 259
    NexusflowDossier

    Arena rating

    Athene V2 Chat, Arena rating: 1,292 (scale 750.0 to 1538.4), 95% confidence interval 1287.0 to 1296.01,292(1287.0–1296.0)
  260. 260
    GoogleDossier

    Arena rating

    Gemma 3 4B, Arena rating: 1,291 (scale 750.0 to 1538.4), 95% confidence interval 1281.6 to 1300.11,291(1281.6–1300.1)
  261. 261

    Arena rating

    Granite 4.1 8B, Arena rating: 1,291 (scale 750.0 to 1538.4), 95% confidence interval 1280.8 to 1300.71,291(1280.8–1300.7)
  262. 262

    Arena rating

    GLM-4-Plus, Arena rating: 1,290 (scale 750.0 to 1538.4), 95% confidence interval 1284.8 to 1294.61,290(1284.8–1294.6)
  263. 263
    TencentDossier

    Arena rating

    Hunyuan Large, Arena rating: 1,288 (scale 750.0 to 1538.4), 95% confidence interval 1278.3 to 1297.71,288(1278.3–1297.7)
  264. 264
    OpenAIDossier

    Arena rating

    gpt-oss-20b, Arena rating: 1,288 (scale 750.0 to 1538.4), 95% confidence interval 1281.2 to 1294.01,288(1281.2–1294.0)
  265. 265

    Arena rating

    Llama 4 Maverick, Arena rating: 1,287 (scale 750.0 to 1538.4), 95% confidence interval 1283.0 to 1291.51,287(1283.0–1291.5)
  266. 266
    Allen Institute for AIDossier

    Arena rating

    Molmo 2 8B, Arena rating: 1,287 (scale 750.0 to 1538.4), 95% confidence interval 1266.3 to 1307.91,287(1266.3–1307.9)
  267. 267
    GoogleDossier

    Arena rating

    Gemini 1.5 Flash (002), Arena rating: 1,287 (scale 750.0 to 1538.4), 95% confidence interval 1282.5 to 1290.91,287(1282.5–1290.9)
  268. 268
    OpenAIDossier

    Arena rating

    GPT-4o mini, Arena rating: 1,286 (scale 750.0 to 1538.4), 95% confidence interval 1282.9 to 1289.91,286(1282.9–1289.9)
  269. 269
    OpenAIDossier

    Arena rating

    GPT-4.1 nano, Arena rating: 1,285 (scale 750.0 to 1538.4), 95% confidence interval 1277.2 to 1292.51,285(1277.2–1292.5)
  270. 270

    Arena rating

    Llama 3.1 405B Instruct (BF16), Arena rating: 1,284 (scale 750.0 to 1538.4), 95% confidence interval 1280.0 to 1287.41,284(1280.0–1287.4)
  271. 271
    NVIDIADossier

    Arena rating

    Llama 3.1 Nemotron 70B Instruct, Arena rating: 1,283 (scale 750.0 to 1538.4), 95% confidence interval 1275.0 to 1290.41,283(1275.0–1290.4)
  272. 272
    OpenAIDossier

    Arena rating

    GPT-4o (Aug '24), Arena rating: 1,283 (scale 750.0 to 1538.4), 95% confidence interval 1278.4 to 1286.81,283(1278.4–1286.8)
  273. 273
    AlibabaDossier

    Arena rating

    Qwen Max (0919), Arena rating: 1,282 (scale 750.0 to 1538.4), 95% confidence interval 1276.3 to 1287.61,282(1276.3–1287.6)
  274. 274

    Arena rating

    Llama 3.1 405B Instruct (FP8), Arena rating: 1,282 (scale 750.0 to 1538.4), 95% confidence interval 1278.3 to 1285.41,282(1278.3–1285.4)
  275. 275
    InceptionDossier

    Arena rating

    Mercury, Arena rating: 1,282 (scale 750.0 to 1538.4), 95% confidence interval 1268.0 to 1295.41,282(1268.0–1295.4)
  276. 276
    AnthropicDossier

    Arena rating

    Claude 3.5 Sonnet (June '24), Arena rating: 1,281 (scale 750.0 to 1538.4), 95% confidence interval 1277.5 to 1284.41,281(1277.5–1284.4)
  277. 277

    Arena rating

    Grok 2 mini, Arena rating: 1,281 (scale 750.0 to 1538.4), 95% confidence interval 1277.0 to 1284.41,281(1277.0–1284.4)
  278. 278

    Arena rating

    Llama 4 Scout, Arena rating: 1,279 (scale 750.0 to 1538.4), 95% confidence interval 1274.6 to 1284.01,279(1274.6–1284.0)
  279. 279
    GoogleDossier

    Arena rating

    Gemini Advanced (0514), Arena rating: 1,279 (scale 750.0 to 1538.4), 95% confidence interval 1273.5 to 1283.71,279(1273.5–1283.7)
  280. 280
    MistralDossier

    Arena rating

    Mistral Small 3.1, Arena rating: 1,277 (scale 750.0 to 1538.4), 95% confidence interval 1272.7 to 1281.81,277(1272.7–1281.8)
  281. 281

    Arena rating

    Llama 3.3 Instruct 70B, Arena rating: 1,274 (scale 750.0 to 1538.4), 95% confidence interval 1270.8 to 1277.81,274(1270.8–1277.8)
  282. 282
    TencentDossier

    Arena rating

    Hunyuan Standard, Arena rating: 1,274 (scale 750.0 to 1538.4), 95% confidence interval 1264.4 to 1283.71,274(1264.4–1283.7)
  283. 283
    GoogleDossier

    Arena rating

    Gemini 1.5 Pro (001), Arena rating: 1,274 (scale 750.0 to 1538.4), 95% confidence interval 1269.6 to 1277.51,274(1269.6–1277.5)
  284. 284
    OpenAIDossier

    Arena rating

    GPT-4 Turbo, Arena rating: 1,272 (scale 750.0 to 1538.4), 95% confidence interval 1267.8 to 1275.61,272(1267.8–1275.6)
  285. 285
    DeepSeekDossier

    Arena rating

    DeepSeek V2.5, Arena rating: 1,271 (scale 750.0 to 1538.4), 95% confidence interval 1266.6 to 1276.01,271(1266.6–1276.0)
  286. 286
    Allen Institute for AIDossier

    Arena rating

    Olmo 3.1 32B Think, Arena rating: 1,271 (scale 750.0 to 1538.4), 95% confidence interval 1263.8 to 1278.31,271(1263.8–1278.3)
  287. 287
    AlibabaDossier

    Arena rating

    Qwen2.5 72B Instruct, Arena rating: 1,269 (scale 750.0 to 1538.4), 95% confidence interval 1265.0 to 1273.11,269(1265.0–1273.1)
  288. 288
    MistralDossier

    Arena rating

    Mistral Large 2 (Jul '24), Arena rating: 1,266 (scale 750.0 to 1538.4), 95% confidence interval 1262.3 to 1270.11,266(1262.3–1270.1)
  289. 289
    MistralDossier

    Arena rating

    Mistral Large (2411), Arena rating: 1,265 (scale 750.0 to 1538.4), 95% confidence interval 1260.9 to 1269.81,265(1260.9–1269.8)
  290. 290
    NexusflowDossier

    Arena rating

    Athene 70B (0725), Arena rating: 1,265 (scale 750.0 to 1538.4), 95% confidence interval 1259.5 to 1270.81,265(1259.5–1270.8)
  291. 291
    OpenAIDossier

    Arena rating

    GPT-4 Turbo Preview (1106), Arena rating: 1,263 (scale 750.0 to 1538.4), 95% confidence interval 1259.6 to 1267.31,263(1259.6–1267.3)
  292. 292
    TencentDossier

    Arena rating

    Hunyuan Large Vision, Arena rating: 1,263 (scale 750.0 to 1538.4), 95% confidence interval 1253.9 to 1272.41,263(1253.9–1272.4)
  293. 293
    OpenAIDossier

    Arena rating

    GPT-4 Turbo Preview (0125), Arena rating: 1,262 (scale 750.0 to 1538.4), 95% confidence interval 1258.2 to 1266.31,262(1258.2–1266.3)
  294. 294
    AnthropicDossier

    Arena rating

    Claude 3 Opus, Arena rating: 1,262 (scale 750.0 to 1538.4), 95% confidence interval 1259.0 to 1265.11,262(1259.0–1265.1)
  295. 295

    Arena rating

    Llama 3.1 Instruct 70B, Arena rating: 1,261 (scale 750.0 to 1538.4), 95% confidence interval 1257.2 to 1264.71,261(1257.2–1264.7)
  296. 296
    AmazonDossier

    Arena rating

    Nova Pro, Arena rating: 1,259 (scale 750.0 to 1538.4), 95% confidence interval 1254.2 to 1263.31,259(1254.2–1263.3)
  297. 297
    Allen Institute for AIDossier

    Arena rating

    Llama 3.1 Tulu 3 70B, Arena rating: 1,256 (scale 750.0 to 1538.4), 95% confidence interval 1245.3 to 1266.21,256(1245.3–1266.2)
  298. 298
    AnthropicDossier

    Arena rating

    Claude 3.5 Haiku, Arena rating: 1,255 (scale 750.0 to 1538.4), 95% confidence interval 1252.2 to 1258.61,255(1252.2–1258.6)
  299. 299
    MistralDossier

    Arena rating

    Magistral Medium (2506), Arena rating: 1,255 (scale 750.0 to 1538.4), 95% confidence interval 1248.3 to 1261.21,255(1248.3–1261.2)
  300. 300
    Reka AIDossier

    Arena rating

    Reka Core, Arena rating: 1,248 (scale 750.0 to 1538.4), 95% confidence interval 1241.3 to 1255.51,248(1241.3–1255.5)
  301. 301

    Arena rating

    Granite 4.0 H Small, Arena rating: 1,241 (scale 750.0 to 1538.4), 95% confidence interval 1233.0 to 1249.91,241(1233.0–1249.9)
  302. 302
    GoogleDossier

    Arena rating

    Gemini 1.5 Flash (001), Arena rating: 1,239 (scale 750.0 to 1538.4), 95% confidence interval 1234.9 to 1243.81,239(1234.9–1243.8)
  303. 303
    AI21 LabsDossier

    Arena rating

    Jamba 1.5 Large, Arena rating: 1,237 (scale 750.0 to 1538.4), 95% confidence interval 1229.8 to 1244.31,237(1229.8–1244.3)
  304. 304
    MistralDossier

    Arena rating

    Mistral Small 3, Arena rating: 1,234 (scale 750.0 to 1538.4), 95% confidence interval 1227.7 to 1239.51,234(1227.7–1239.5)
  305. 305
    GoogleDossier

    Arena rating

    Gemma 2 27B IT, Arena rating: 1,231 (scale 750.0 to 1538.4), 95% confidence interval 1228.1 to 1234.81,231(1228.1–1234.8)
  306. 306
    AlibabaDossier

    Arena rating

    Qwen2.5 Coder 32B Instruct, Arena rating: 1,230 (scale 750.0 to 1538.4), 95% confidence interval 1222.0 to 1238.11,230(1222.0–1238.1)
  307. 307
    CohereDossier

    Arena rating

    Command R+ (08-2024), Arena rating: 1,229 (scale 750.0 to 1538.4), 95% confidence interval 1222.4 to 1235.51,229(1222.4–1235.5)
  308. 308
    AmazonDossier

    Arena rating

    Nova Lite, Arena rating: 1,229 (scale 750.0 to 1538.4), 95% confidence interval 1223.4 to 1233.71,229(1223.4–1233.7)
  309. 309
    NVIDIADossier

    Arena rating

    Llama 3.1 Nemotron 51B Instruct, Arena rating: 1,228 (scale 750.0 to 1538.4), 95% confidence interval 1218.1 to 1238.01,228(1218.1–1238.0)
  310. 310
    Princeton NLPDossier

    Arena rating

    Gemma 2 9B IT SimPO, Arena rating: 1,227 (scale 750.0 to 1538.4), 95% confidence interval 1220.3 to 1233.91,227(1220.3–1233.9)
  311. 311

    Arena rating

    GLM-4 (0520), Arena rating: 1,226 (scale 750.0 to 1538.4), 95% confidence interval 1219.0 to 1232.91,226(1219.0–1232.9)
  312. 312
    GoogleDossier

    Arena rating

    Gemini 1.5 Flash-8B (001), Arena rating: 1,226 (scale 750.0 to 1538.4), 95% confidence interval 1221.6 to 1230.21,226(1221.6–1230.2)
  313. 313
    NVIDIADossier

    Arena rating

    Nemotron-4 340B Instruct, Arena rating: 1,225 (scale 750.0 to 1538.4), 95% confidence interval 1219.6 to 1230.11,225(1219.6–1230.1)
  314. 314
    CohereDossier

    Arena rating

    Aya Expanse 32B, Arena rating: 1,224 (scale 750.0 to 1538.4), 95% confidence interval 1219.2 to 1228.91,224(1219.2–1228.9)
  315. 315

    Arena rating

    Llama 3 Instruct 70B, Arena rating: 1,221 (scale 750.0 to 1538.4), 95% confidence interval 1217.3 to 1224.41,221(1217.3–1224.4)
  316. 316
    AnthropicDossier

    Arena rating

    Claude 3 Sonnet, Arena rating: 1,218 (scale 750.0 to 1538.4), 95% confidence interval 1214.1 to 1222.01,218(1214.1–1222.0)
  317. 317
    Allen Institute for AIDossier

    Arena rating

    OLMo 2 32B Instruct (0325), Arena rating: 1,218 (scale 750.0 to 1538.4), 95% confidence interval 1207.2 to 1228.81,218(1207.2–1228.8)
  318. 318
    Reka AIDossier

    Arena rating

    Reka Flash (Sep '24), Arena rating: 1,218 (scale 750.0 to 1538.4), 95% confidence interval 1211.0 to 1224.91,218(1211.0–1224.9)
  319. 319
    MicrosoftDossier

    Arena rating

    Phi-4, Arena rating: 1,217 (scale 750.0 to 1538.4), 95% confidence interval 1212.2 to 1221.31,217(1212.2–1221.3)
  320. 320
    AmazonDossier

    Arena rating

    Nova Micro, Arena rating: 1,208 (scale 750.0 to 1538.4), 95% confidence interval 1203.3 to 1213.61,208(1203.3–1213.6)
  321. 321
    GoogleDossier

    Arena rating

    Gemma 2 9B IT, Arena rating: 1,207 (scale 750.0 to 1538.4), 95% confidence interval 1203.6 to 1211.21,207(1203.6–1211.2)
  322. 322
    OpenAIDossier

    Arena rating

    GPT-4 (0314), Arena rating: 1,206 (scale 750.0 to 1538.4), 95% confidence interval 1201.3 to 1210.81,206(1201.3–1210.8)
  323. 323
    CohereDossier

    Arena rating

    Command R+, Arena rating: 1,204 (scale 750.0 to 1538.4), 95% confidence interval 1199.6 to 1208.21,204(1199.6–1208.2)
  324. 324
    AlibabaDossier

    Arena rating

    Qwen2 72B Instruct, Arena rating: 1,203 (scale 750.0 to 1538.4), 95% confidence interval 1198.4 to 1208.21,203(1198.4–1208.2)
  325. 325
    TencentDossier

    Arena rating

    Hunyuan Standard 256K, Arena rating: 1,202 (scale 750.0 to 1538.4), 95% confidence interval 1190.4 to 1213.51,202(1190.4–1213.5)
  326. 326
    AnthropicDossier

    Arena rating

    Claude 3 Haiku, Arena rating: 1,195 (scale 750.0 to 1538.4), 95% confidence interval 1191.0 to 1198.51,195(1191.0–1198.5)
  327. 327
    Allen Institute for AIDossier

    Arena rating

    Llama 3.1 Tulu 3 8B, Arena rating: 1,193 (scale 750.0 to 1538.4), 95% confidence interval 1182.9 to 1203.81,193(1182.9–1203.8)
  328. 328
    DeepSeekDossier

    Arena rating

    DeepSeek Coder V2, Arena rating: 1,191 (scale 750.0 to 1538.4), 95% confidence interval 1184.9 to 1197.61,191(1184.9–1197.6)
  329. 329
    MistralDossier

    Arena rating

    Ministral 8B (2410), Arena rating: 1,191 (scale 750.0 to 1538.4), 95% confidence interval 1182.0 to 1199.91,191(1182.0–1199.9)
  330. 330
    CohereDossier

    Arena rating

    Command R (08-2024), Arena rating: 1,187 (scale 750.0 to 1538.4), 95% confidence interval 1180.9 to 1193.91,187(1180.9–1193.9)
  331. 331
    AI21 LabsDossier

    Arena rating

    Jamba 1.5 Mini, Arena rating: 1,187 (scale 750.0 to 1538.4), 95% confidence interval 1179.6 to 1193.91,187(1179.6–1193.9)
  332. 332

    Arena rating

    Llama 3.1 Instruct 8B, Arena rating: 1,187 (scale 750.0 to 1538.4), 95% confidence interval 1182.5 to 1190.71,187(1182.5–1190.7)
  333. 333
    OpenAIDossier

    Arena rating

    GPT-4 (0613), Arena rating: 1,186 (scale 750.0 to 1538.4), 95% confidence interval 1182.0 to 1190.11,186(1182.0–1190.1)
  334. 334
    CohereDossier

    Arena rating

    Aya Expanse 8B, Arena rating: 1,185 (scale 750.0 to 1538.4), 95% confidence interval 1178.2 to 1192.11,185(1178.2–1192.1)
  335. 335
    MistralDossier

    Arena rating

    Mistral Large (Feb '24), Arena rating: 1,177 (scale 750.0 to 1538.4), 95% confidence interval 1171.8 to 1181.21,177(1171.8–1181.2)
  336. 336
    AlibabaDossier

    Arena rating

    Qwen1.5 110B Chat, Arena rating: 1,175 (scale 750.0 to 1538.4), 95% confidence interval 1169.1 to 1180.11,175(1169.1–1180.1)
  337. 337
    01.AIDossier

    Arena rating

    Yi 1.5 34B Chat, Arena rating: 1,173 (scale 750.0 to 1538.4), 95% confidence interval 1168.0 to 1178.01,173(1168.0–1178.0)
  338. 338
    Reka AIDossier

    online

    Arena rating

    Reka Flash 21B, Arena rating: 1,170 (scale 750.0 to 1538.4), 95% confidence interval 1163.2 to 1177.81,170(1163.2–1177.8)
  339. 339
    AlibabaDossier

    Arena rating

    Qwen1.5 72B Chat, Arena rating: 1,166 (scale 750.0 to 1538.4), 95% confidence interval 1161.0 to 1171.41,166(1161.0–1171.4)
  340. 340

    Arena rating

    Llama 3 Instruct 8B, Arena rating: 1,166 (scale 750.0 to 1538.4), 95% confidence interval 1161.9 to 1169.41,166(1161.9–1169.4)
  341. 341
    MistralDossier

    Arena rating

    Mistral Medium, Arena rating: 1,165 (scale 750.0 to 1538.4), 95% confidence interval 1159.8 to 1170.61,165(1159.8–1170.6)
  342. 342
    Reka AIDossier

    Arena rating

    Reka Flash 21B, Arena rating: 1,165 (scale 750.0 to 1538.4), 95% confidence interval 1159.1 to 1170.91,165(1159.1–1170.9)
  343. 343
    CohereDossier

    Arena rating

    Command R, Arena rating: 1,163 (scale 750.0 to 1538.4), 95% confidence interval 1158.4 to 1167.91,163(1158.4–1167.9)
  344. 344
    MistralDossier

    Arena rating

    Mixtral 8x22B Instruct v0.1, Arena rating: 1,162 (scale 750.0 to 1538.4), 95% confidence interval 1157.5 to 1166.51,162(1157.5–1166.5)
  345. 345
    AlibabaDossier

    Arena rating

    QwQ 32B Preview, Arena rating: 1,162 (scale 750.0 to 1538.4), 95% confidence interval 1150.5 to 1173.21,162(1150.5–1173.2)
  346. 346
    Shanghai AI LaboratoryDossier

    Arena rating

    InternLM2.5 20B Chat, Arena rating: 1,158 (scale 750.0 to 1538.4), 95% confidence interval 1151.4 to 1165.51,158(1151.4–1165.5)
  347. 347
    GoogleDossier

    Arena rating

    Gemma 2 2B IT, Arena rating: 1,156 (scale 750.0 to 1538.4), 95% confidence interval 1151.8 to 1159.91,156(1151.8–1159.9)
  348. 348

    Arena rating

    Granite 3.1 8B Instruct, Arena rating: 1,150 (scale 750.0 to 1538.4), 95% confidence interval 1138.6 to 1160.51,150(1138.6–1160.5)
  349. 349
    GoogleDossier

    dev-api

    Arena rating

    Gemini Pro, Arena rating: 1,149 (scale 750.0 to 1538.4), 95% confidence interval 1141.7 to 1156.11,149(1141.7–1156.1)
  350. 350
    Hugging FaceDossier

    Arena rating

    Zephyr ORPO 141B A35B v0.1, Arena rating: 1,144 (scale 750.0 to 1538.4), 95% confidence interval 1133.0 to 1154.41,144(1133.0–1154.4)
  351. 351
    MicrosoftDossier

    Arena rating

    Phi-3 Medium 4K Instruct, Arena rating: 1,138 (scale 750.0 to 1538.4), 95% confidence interval 1132.5 to 1142.71,138(1132.5–1142.7)
  352. 352
    AlibabaDossier

    Arena rating

    Qwen1.5 32B Chat, Arena rating: 1,137 (scale 750.0 to 1538.4), 95% confidence interval 1131.2 to 1143.31,137(1131.2–1143.3)
  353. 353
    NexusflowDossier

    Arena rating

    Starling LM 7B Beta, Arena rating: 1,132 (scale 750.0 to 1538.4), 95% confidence interval 1124.8 to 1139.41,132(1124.8–1139.4)
  354. 354
    MistralDossier

    Arena rating

    Mixtral 8x7B Instruct, Arena rating: 1,132 (scale 750.0 to 1538.4), 95% confidence interval 1127.4 to 1135.81,132(1127.4–1135.8)
  355. 355
    GoogleDossier

    Arena rating

    Gemini Pro, Arena rating: 1,131 (scale 750.0 to 1538.4), 95% confidence interval 1119.2 to 1142.01,131(1119.2–1142.0)
  356. 356
    01.AIDossier

    Arena rating

    Yi 34B Chat, Arena rating: 1,129 (scale 750.0 to 1538.4), 95% confidence interval 1121.7 to 1135.31,129(1121.7–1135.3)
  357. 357
    AlibabaDossier

    Arena rating

    Qwen1.5 14B Chat, Arena rating: 1,128 (scale 750.0 to 1538.4), 95% confidence interval 1120.7 to 1134.81,128(1120.7–1134.8)
  358. 358

    Arena rating

    Granite 3.1 2B Instruct, Arena rating: 1,127 (scale 750.0 to 1538.4), 95% confidence interval 1116.4 to 1138.51,127(1116.4–1138.5)
  359. 359
    OpenAIDossier

    Arena rating

    GPT-3.5 Turbo (0125), Arena rating: 1,125 (scale 750.0 to 1538.4), 95% confidence interval 1120.3 to 1129.61,125(1120.3–1129.6)
  360. 360
    Allen Institute for AIDossier

    Arena rating

    Tulu 2 DPO 70B, Arena rating: 1,121 (scale 750.0 to 1538.4), 95% confidence interval 1111.4 to 1130.81,121(1111.4–1130.8)
  361. 361
    MicrosoftDossier

    Arena rating

    WizardLM 70B, Arena rating: 1,120 (scale 750.0 to 1538.4), 95% confidence interval 1110.2 to 1128.91,120(1110.2–1128.9)
  362. 362
    DatabricksDossier

    Arena rating

    DBRX Instruct Preview, Arena rating: 1,119 (scale 750.0 to 1538.4), 95% confidence interval 1112.9 to 1125.01,119(1112.9–1125.0)
  363. 363

    Arena rating

    Llama 2 70B Chat, Arena rating: 1,115 (scale 750.0 to 1538.4), 95% confidence interval 1109.8 to 1120.71,115(1109.8–1120.7)
  364. 364
    Nous ResearchDossier

    Arena rating

    Nous Hermes 2 Mixtral 8x7B DPO, Arena rating: 1,112 (scale 750.0 to 1538.4), 95% confidence interval 1100.3 to 1123.91,112(1100.3–1123.9)
  365. 365
    MicrosoftDossier

    Arena rating

    Phi-3 Small 8K Instruct, Arena rating: 1,110 (scale 750.0 to 1538.4), 95% confidence interval 1104.0 to 1115.71,110(1104.0–1115.7)
  366. 366

    Arena rating

    Llama 3.2 3B Instruct, Arena rating: 1,110 (scale 750.0 to 1538.4), 95% confidence interval 1102.0 to 1117.11,110(1102.0–1117.1)
  367. 367
    UC BerkeleyDossier

    Arena rating

    Starling LM 7B Alpha, Arena rating: 1,108 (scale 750.0 to 1538.4), 95% confidence interval 1099.8 to 1115.61,108(1099.8–1115.6)
  368. 368
    OpenChatDossier

    Arena rating

    OpenChat 3.5 (0106), Arena rating: 1,107 (scale 750.0 to 1538.4), 95% confidence interval 1098.8 to 1114.51,107(1098.8–1114.5)
  369. 369
    LMSYSDossier

    Arena rating

    Vicuna 33B, Arena rating: 1,105 (scale 750.0 to 1538.4), 95% confidence interval 1099.2 to 1111.41,105(1099.2–1111.4)
  370. 370
    DeepSeekDossier

    Arena rating

    DeepSeek LLM 67B Chat, Arena rating: 1,105 (scale 750.0 to 1538.4), 95% confidence interval 1093.5 to 1116.61,105(1093.5–1116.6)
  371. 371
    SnowflakeDossier

    Arena rating

    Snowflake Arctic Instruct, Arena rating: 1,101 (scale 750.0 to 1538.4), 95% confidence interval 1095.1 to 1106.91,101(1095.1–1106.9)
  372. 372
    NVIDIADossier

    Arena rating

    Llama 2 70B SteerLM Chat, Arena rating: 1,098 (scale 750.0 to 1538.4), 95% confidence interval 1085.7 to 1110.61,098(1085.7–1110.6)
  373. 373
    OpenChatDossier

    Arena rating

    OpenChat 3.5, Arena rating: 1,097 (scale 750.0 to 1538.4), 95% confidence interval 1087.5 to 1106.71,097(1087.5–1106.7)
  374. 374

    Arena rating

    Granite 3.0 8B Instruct, Arena rating: 1,096 (scale 750.0 to 1538.4), 95% confidence interval 1088.0 to 1105.01,096(1088.0–1105.0)
  375. 375
    OpenAIDossier

    Arena rating

    GPT-3.5 Turbo (1106), Arena rating: 1,094 (scale 750.0 to 1538.4), 95% confidence interval 1085.4 to 1102.91,094(1085.4–1102.9)
  376. 376
    GoogleDossier

    Arena rating

    Gemma 1.1 7B IT, Arena rating: 1,094 (scale 750.0 to 1538.4), 95% confidence interval 1088.0 to 1100.01,094(1088.0–1100.0)
  377. 377
    TekniumDossier

    Arena rating

    OpenHermes 2.5 Mistral 7B, Arena rating: 1,093 (scale 750.0 to 1538.4), 95% confidence interval 1082.2 to 1102.81,093(1082.2–1102.8)
  378. 378
    MistralDossier

    Arena rating

    Mistral 7B Instruct v0.2, Arena rating: 1,090 (scale 750.0 to 1538.4), 95% confidence interval 1083.1 to 1096.11,090(1083.1–1096.1)
  379. 379

    Arena rating

    Llama 2 13B Chat, Arena rating: 1,084 (scale 750.0 to 1538.4), 95% confidence interval 1077.8 to 1091.21,084(1077.8–1091.2)
  380. 380
    AlibabaDossier

    Arena rating

    Qwen1.5 7B Chat, Arena rating: 1,084 (scale 750.0 to 1538.4), 95% confidence interval 1073.8 to 1093.21,084(1073.8–1093.2)
  381. 381
    UpstageDossier

    Arena rating

    SOLAR 10.7B Instruct v1.0, Arena rating: 1,083 (scale 750.0 to 1538.4), 95% confidence interval 1070.1 to 1096.21,083(1070.1–1096.2)
  382. 382
    Cognitive ComputationsDossier

    Arena rating

    Dolphin 2.2.1 Mistral 7B, Arena rating: 1,081 (scale 750.0 to 1538.4), 95% confidence interval 1065.7 to 1096.31,081(1065.7–1096.3)
  383. 383
    MicrosoftDossier

    Arena rating

    Phi-3 Mini 4K Instruct (June 2024), Arena rating: 1,080 (scale 750.0 to 1538.4), 95% confidence interval 1073.7 to 1086.31,080(1073.7–1086.3)
  384. 384

    Arena rating

    Granite 3.0 2B Instruct, Arena rating: 1,080 (scale 750.0 to 1538.4), 95% confidence interval 1071.7 to 1088.11,080(1071.7–1088.1)
  385. 385
    MicrosoftDossier

    Arena rating

    WizardLM 13B, Arena rating: 1,077 (scale 750.0 to 1538.4), 95% confidence interval 1067.4 to 1085.81,077(1067.4–1085.8)
  386. 386
    MicrosoftDossier

    Arena rating

    Phi-3 Mini 4K Instruct, Arena rating: 1,073 (scale 750.0 to 1538.4), 95% confidence interval 1067.1 to 1079.61,073(1067.1–1079.6)
  387. 387
    Hugging FaceDossier

    Arena rating

    Zephyr 7B Beta, Arena rating: 1,069 (scale 750.0 to 1538.4), 95% confidence interval 1060.8 to 1078.21,069(1060.8–1078.2)
  388. 388
    MosaicMLDossier

    Arena rating

    MPT 30B Chat, Arena rating: 1,069 (scale 750.0 to 1538.4), 95% confidence interval 1056.6 to 1080.71,069(1056.6–1080.7)
  389. 389

    Arena rating

    Code Llama 34B Instruct, Arena rating: 1,066 (scale 750.0 to 1538.4), 95% confidence interval 1056.8 to 1074.31,066(1056.8–1074.3)
  390. 390
    Hugging FaceDossier

    Arena rating

    Zephyr 7B Alpha, Arena rating: 1,058 (scale 750.0 to 1538.4), 95% confidence interval 1042.8 to 1074.11,058(1042.8–1074.1)
  391. 391
    LMSYSDossier

    Arena rating

    Vicuna 13B, Arena rating: 1,058 (scale 750.0 to 1538.4), 95% confidence interval 1051.1 to 1064.31,058(1051.1–1064.3)
  392. 392

    Arena rating

    Code Llama 70B Instruct, Arena rating: 1,057 (scale 750.0 to 1538.4), 95% confidence interval 1039.3 to 1075.31,057(1039.3–1075.3)
  393. 393
    GoogleDossier

    Arena rating

    Gemma 7B IT, Arena rating: 1,056 (scale 750.0 to 1538.4), 95% confidence interval 1046.7 to 1065.71,056(1046.7–1065.7)
  394. 394

    Arena rating

    Llama 3.2 1B Instruct, Arena rating: 1,055 (scale 750.0 to 1538.4), 95% confidence interval 1047.0 to 1062.41,055(1047.0–1062.4)
  395. 395
    Technology Innovation InstituteDossier

    Arena rating

    Falcon 180B Chat, Arena rating: 1,054 (scale 750.0 to 1538.4), 95% confidence interval 1037.4 to 1071.41,054(1037.4–1071.4)
  396. 396

    Arena rating

    Llama 2 7B Chat, Arena rating: 1,053 (scale 750.0 to 1538.4), 95% confidence interval 1046.5 to 1060.41,053(1046.5–1060.4)
  397. 397
    University of WashingtonDossier

    Arena rating

    Guanaco 33B, Arena rating: 1,053 (scale 750.0 to 1538.4), 95% confidence interval 1041.4 to 1065.41,053(1041.4–1065.4)
  398. 398
    AlibabaDossier

    Arena rating

    Qwen 14B Chat, Arena rating: 1,051 (scale 750.0 to 1538.4), 95% confidence interval 1040.2 to 1061.91,051(1040.2–1061.9)
  399. 399
    MicrosoftDossier

    Arena rating

    Phi-3 Mini 128K Instruct, Arena rating: 1,050 (scale 750.0 to 1538.4), 95% confidence interval 1043.2 to 1057.61,050(1043.2–1057.6)
  400. 400
    Hugging FaceDossier

    Arena rating

    SmolLM2 1.7B Instruct, Arena rating: 1,042 (scale 750.0 to 1538.4), 95% confidence interval 1028.0 to 1056.11,042(1028.0–1056.1)
  401. 401
    Together AIDossier

    Arena rating

    StripedHyena Nous 7B, Arena rating: 1,038 (scale 750.0 to 1538.4), 95% confidence interval 1027.5 to 1049.41,038(1027.5–1049.4)
  402. 402
    Allen Institute for AIDossier

    Arena rating

    OLMo 7B Instruct, Arena rating: 1,032 (scale 750.0 to 1538.4), 95% confidence interval 1020.8 to 1042.91,032(1020.8–1042.9)
  403. 403
    LMSYSDossier

    Arena rating

    Vicuna 7B, Arena rating: 1,031 (scale 750.0 to 1538.4), 95% confidence interval 1021.9 to 1040.11,031(1021.9–1040.1)
  404. 404
    GoogleDossier

    Arena rating

    PaLM 2, Arena rating: 1,027 (scale 750.0 to 1538.4), 95% confidence interval 1018.1 to 1036.51,027(1018.1–1036.5)
  405. 405
    MistralDossier

    Arena rating

    Mistral 7B Instruct, Arena rating: 1,024 (scale 750.0 to 1538.4), 95% confidence interval 1014.4 to 1032.71,024(1014.4–1032.7)
  406. 406
    GoogleDossier

    Arena rating

    Gemma 1.1 2B IT, Arena rating: 1,022 (scale 750.0 to 1538.4), 95% confidence interval 1014.6 to 1029.91,022(1014.6–1029.9)
  407. 407
    GoogleDossier

    Arena rating

    Gemma 2B IT, Arena rating: 1,002 (scale 750.0 to 1538.4), 95% confidence interval 990.7 to 1013.71,002(990.7–1013.7)
  408. 408
    AlibabaDossier

    Arena rating

    Qwen1.5 4B Chat, Arena rating: 997 (scale 750.0 to 1538.4), 95% confidence interval 988.1 to 1006.5997(988.1–1006.5)
  409. 409
    UC BerkeleyDossier

    Arena rating

    Koala 13B, Arena rating: 990 (scale 750.0 to 1538.4), 95% confidence interval 980.0 to 999.3990(980.0–999.3)
  410. 410

    Arena rating

    ChatGLM3 6B, Arena rating: 972 (scale 750.0 to 1538.4), 95% confidence interval 960.5 to 983.6972(960.5–983.6)
  411. 411
    Nomic AIDossier

    Arena rating

    GPT4All 13B Snoozy, Arena rating: 956 (scale 750.0 to 1538.4), 95% confidence interval 941.2 to 971.2956(941.2–971.2)
  412. 412
    MosaicMLDossier

    Arena rating

    MPT 7B Chat, Arena rating: 955 (scale 750.0 to 1538.4), 95% confidence interval 943.6 to 967.0955(943.6–967.0)
  413. 413

    Arena rating

    RWKV-4 Raven 14B, Arena rating: 948 (scale 750.0 to 1538.4), 95% confidence interval 937.2 to 959.7948(937.2–959.7)
  414. 414

    Arena rating

    ChatGLM2 6B, Arena rating: 939 (scale 750.0 to 1538.4), 95% confidence interval 925.2 to 952.1939(925.2–952.1)
  415. 415
    Stanford UniversityDossier

    Arena rating

    Alpaca 13B, Arena rating: 933 (scale 750.0 to 1538.4), 95% confidence interval 921.6 to 944.1933(921.6–944.1)
  416. 416

    Arena rating

    ChatGLM 6B, Arena rating: 919 (scale 750.0 to 1538.4), 95% confidence interval 906.3 to 930.7919(906.3–930.7)
  417. 417
    Open-Assistant (LAION)Dossier

    Arena rating

    OpenAssistant Pythia 12B, Arena rating: 916 (scale 750.0 to 1538.4), 95% confidence interval 905.4 to 926.8916(905.4–926.8)
  418. 418
    LMSYSDossier

    Arena rating

    FastChat-T5 3B, Arena rating: 894 (scale 750.0 to 1538.4), 95% confidence interval 882.2 to 906.6894(882.2–906.6)
  419. 419
    Stability AIDossier

    Arena rating

    StableLM Tuned Alpha 7B, Arena rating: 867 (scale 750.0 to 1538.4), 95% confidence interval 854.1 to 879.2867(854.1–879.2)
  420. 420
    DatabricksDossier

    Arena rating

    Dolly 2.0 12B, Arena rating: 851 (scale 750.0 to 1538.4), 95% confidence interval 837.9 to 864.4851(837.9–864.4)
  421. 421

    Arena rating

    LLaMA 13B, Arena rating: 834 (scale 750.0 to 1538.4), 95% confidence interval 818.0 to 849.1834(818.0–849.1)