Boards

Ranked by a named score.

Voice has no licensed leaderboard in this snapshot, so it is not given a made-up ranking. Open the directory and filter to voice.

Reasoning

LMArena text overall rating. Higher is the board's own Elo-style rating. LMArena dataset is CC-BY-4.0. Epoch is CC-BY-4.0.

RankModelScoreVotes
1Gemini 4 Argon (High) google1532.34,900
2Claude Opus 5.5 (High) anthropic1516.29,496
3Claude Fable 5.1 (Max) anthropic1511.915,623
4Claude Opus 5 (Max) anthropic1508.633,683
5Claude Opus 5 (High) anthropic1502.869,307
6Muse Wonder (Max) meta1501.75,099
7Claude Opus 4.6 (High) anthropic1501.282,619
8Gemini 3.8 Flash (High) google1498.941,423
9Claude Opus 4.6 anthropic1497.087,372
10Claude Fable 5 (High) anthropic1491.343,399
11MiMo-V2.6-Pro xiaomi1491.14,065
12Muse Spark 1.3 (Max) meta1490.519,323
13Claude Opus 4.7 (High) anthropic1489.765,131
14Gemini 3.7 Flash (High) google1488.529,177
15Qwen3.8-Max alibaba1483.229,418
16Claude Opus 4.7 anthropic1483.066,185
17Muse Spark 1.2 (xHigh) meta1482.49,578
18Gemini 3.5 Flash (High) google1482.048,479
19Gemini 3.1 Pro Preview google1480.1129,523
20Gemini 3.6 Flash (High) google1479.536,079
21Gemini 3 Pro google1479.542,059
22Muse Spark 1.1 meta1478.344,815
23Gemini 3.5 Flash (Medium) google1477.647,228
24Kimi K3 (Max) moonshot1475.734,325
25Qwen3.7-Max-Preview alibaba1474.93,942

Coding

LMArena text category 'coding'. LMArena dataset is CC-BY-4.0. Epoch is CC-BY-4.0.

RankModelScoreVotes
1Gemini 4 Argon (High) google1546.01,236
2Claude Opus 5.5 (High) anthropic1545.02,143
3Claude Opus 4.6 anthropic1534.524,104
4Claude Opus 5 (Max) anthropic1533.28,413
5MiMo-V2.6-Pro xiaomi1533.21,102
6Claude Fable 5.1 (Max) anthropic1532.53,112
7Claude Opus 5 (High) anthropic1532.317,823
8Claude Opus 4.6 (High) anthropic1530.721,348
9Muse Wonder (Max) meta1522.11,234
10Claude Opus 4.7 (High) anthropic1518.518,580
11Claude Fable 5 (High) anthropic1518.311,128
12Claude Opus 4.7 anthropic1514.818,786
13Claude Sonnet 5.5 (xHigh) anthropic1513.11,914
14Muse Spark 1.3 (Max) meta1510.24,958
15Gemini 3.8 Flash (High) google1508.310,731
16Kimi K3 (Max) moonshot1508.38,643
17GLM-5.3-Flash zai1506.89,192
18Muse Spark 1.2 (xHigh) meta1505.22,474
19Claude Opus 4.5 (High, 32K, 20251101) anthropic1503.57,787
20Claude Sonnet 4.6 anthropic1503.519,732
21DeepSeek-V4.1-Flash-20260910 (Max) deepseek1502.13,680
22Qwen3.8-Max alibaba1501.68,078
23MiMo-V2.5-Pro xiaomi1501.120,773
24MiMo-V2.6-Flash xiaomi1499.13,010
25Gemini 3.7 Flash (High) google1498.97,905

Math

LMArena text category 'math'. LMArena dataset is CC-BY-4.0. Epoch is CC-BY-4.0.

RankModelScoreVotes
1Gemini 4 Argon (High) google1536.4273
2Claude Opus 5 (Max) anthropic1533.11,597
3Gemini 3.8 Flash (High) google1532.02,094
4Claude Opus 5 (High) anthropic1530.93,315
5Claude Fable 5.1 (Max) anthropic1530.8746
6Claude Opus 5.5 (High) anthropic1517.7491
7Claude Opus 4.6 (High) anthropic1517.64,248
8Claude Fable 5 (High) anthropic1516.52,161
9Gemini 3.7 Flash (High) google1510.91,501
10Claude Opus 4.6 anthropic1509.54,741
11Gemini 3.6 Flash (High) google1507.51,760
12Gemini 3.5 Flash (High) google1504.22,388
13GLM-5.3-Flash zai1502.71,820
14Qwen3.8-Max alibaba1501.41,481
15MiMo-V2.6-Pro xiaomi1499.9215
16Claude Opus 4.7 (High) anthropic1498.53,293
17Muse Spark 1.3 (Max) meta1496.9948
18Claude Sonnet 5.5 (xHigh) anthropic1496.7416
19GLM-5.3 (Max) zai1495.41,023
20Kimi K3 (Max) moonshot1493.81,535
21Claude Opus 4.7 anthropic1489.33,389
22GPT-6.1 Sol (Max) openai1488.9390
23Qwen3.7-Max-Preview alibaba1487.9233
24GPT-5.4 (High) openai1487.53,377
25GPT-5.5 openai1487.03,518

Image

LMArena text-to-image board. LMArena dataset is CC-BY-4.0. Epoch is CC-BY-4.0.

RankModelScoreVotes
1gpt-image-2.5-sunburst openai1424.517,752
2gpt-image-2.5-flare openai1397.616,889
3gpt-image-2 (medium) openai1383.093,334
4grok-imagine-image-2.0 (canvas) xai1333.53,113
5mai-image-2.6 microsoft-ai1332.125,091
6gemini-nano-banana-2.1 google1327.85,310
7reve-2.1 reve1301.78,249
8grok-imagine-image-2.0 (20260801) xai1294.214,483
9muse-image meta1274.042,262
10reve-2.0 reve1269.215,773
11gemini-3.1-flash-image (nano-banana-2) [web-search] google1261.860,515
12seedream-5.0-pro bytedance1255.594,181
13qwen-image-3.0-pro alibaba1255.211,440
14mai-image-2.5 microsoft-ai1254.261,265
15gemini-3.1-flash-lite-image (nano-banana-2-lite) google1250.217,804
16gemini-3-pro-image-2k (nano-banana-pro) google1247.9175,734
17gpt-image-1.5-high-fidelity openai1237.4169,927
18gemini-3-pro-image-preview (nano-banana-pro) google1234.186,464
19qwen-image-2.1 alibaba1222.911,441
20ideogram-4.0-quality ideogram1205.160,591
21qwen-image-2.0-pro-2026-06-22 alibaba1191.813,096
22uni-1.1-max luma-ai1189.314,072
23uni-1.1 luma-ai1184.346,699
24mai-image-2 microsoft-ai1183.851,475
25grok-imagine-image xai1169.7272,085

Video

LMArena image-to-video board. LMArena dataset is CC-BY-4.0. Epoch is CC-BY-4.0.

RankModelScoreVotes
1minimax-h3 minimax1495.457,112
2gemini-omni-1.1-flash google1488.53,734
3wan3.0 alibaba1479.83,442
4dreamina-seedance-2.5-720p bytedance1477.415,729
5dreamina-seedance-2.0-720p bytedance1474.6121,427
6gemini-omni-flash google1464.887,155
7hidream-o1-video-1.0 hidream1456.56,488
8grok-imagine-video-1.5-720p xai1456.0137,459
9flux-3-video-20260811 bfl1449.537,042
10happyhorse-1.0 aorizon1441.770,640
11wan2.7-i2v wan1428.285,521
12grok-imagine-video-720p xai1415.3554,050
13veo-3.1-audio google1398.125,116
14veo-3.1-audio-1080p google1390.253,089
15veo-3.1-fast-audio google1385.199,804
16grok-imagine-video-480p xai1384.919,429
17veo-3.1-fast-audio-1080p google1371.654,380
18vidu-q3-pro shengshu1363.236,679
19kling-v3-pro kling1354.0206,221
20veo-3-audio google1331.532,367
21veo-3-fast-audio google1326.041,194
22wan2.5-i2v-preview alibaba1321.416,907
23wan2.6-i2v alibaba1309.5121,082
24seedance-v1.5-pro bytedance1307.2245,206
25pixverse-v5.6 1299.3126,710

GPQA Diamond

Epoch AI, percent. CC-BY-4.0. Not mixed into the Value Index. LMArena dataset is CC-BY-4.0. Epoch is CC-BY-4.0.

RankModelScoreVotes
1Gemini 4 Argon Google96.4%—
2GPT-6 Astra OpenAI95.8%—
3Claude Sonnet 5.5 Anthropic95.6%—
4GPT-6.1 Sol OpenAI95.4%—
5Gemini 3.8 Flash Google95.4%—
6Gemini 3.7 Flash Google94.8%—
7GPT-6 Sol OpenAI94.3%—
8Gemini 3.6 Flash Google94.1%—
9Gemini 3.1 Pro Preview Google94.1%—
10Grok 4.6 xAI94.0%—
11Claude Opus 5 Anthropic93.9%—
12GPT-5.6 Sol OpenAI93.5%—
13Grok 4.5 xAI93.4%—
14GPT-5.6 Terra OpenAI93.3%—
15DeepSeek V4.1 Flash 1017 DeepSeek93.2%—
16Kimi K3 Nvidia93.1%—
17Gemini 3.5 Flash Google92.8%—
18Grok 4.7 xAI92.7%—
19Qwen3.8 Max Alibaba92.7%—
20GLM-5.2 Alibaba91.9%—
21GPT-5.6 Luna OpenAI91.6%—
22Claude Opus 4.8 Anthropic91.0%—
23DeepSeek V4 Flash 0731 Alibaba91.0%—
24GLM-5.3 Nvidia90.9%—
25Qwen3.7 Max Alibaba90.9%—