Leaderboard

Human-voted ELO, computed from the votes on all documents · updates live with every vote.

1 combo is hidden until they have enough votes to rank reliably (at least 174 in this view).

#ModelTodayScoreGamesWin %
1GPT-5.6 Sol=
1294
48879%
2Grok 4.5=
1243
47475%
3Qwen3.7 Plus=
1231
49174%
4Gemini 3.6 Flash=
1224
42774%
5MiniMax-M3=
1223
46473%
6Qwen3.7 Max=
1197
49570%
7Claude Opus 4.8=
1193
46569%
8Kimi K3=
1176
44668%
9GLM-5.2=
1167
46167%
10Claude Opus 5=
1162
21166%
11GPT-5.6 Terra=
1147
47964%
12DeepSeek V4 Flash=
1128
46363%
13Claude Fable 5=
1116
46461%
14Gemini 3.5 Flash=
1112
45761%
15DeepSeek V4 Pro=
1101
45960%
16Claude Sonnet 5=
1090
46858%
17GPT-5.6 Luna=
1074
47854%
18Qwen3.7 Flash=
1052
20653%
19Muse Spark 1.1=
1017
45551%
20KAT-Coder-Pro V2.5=
1001
43848%
21GPT-5.5=
999
45548%
22Kimi K2.7 Code=
980
46645%
23Nex N2 Pro=
977
44146%
24Inkling=
956
43943%
25Step 3.7 Flash=
860
45633%
26Qwen3.5 397B A17B=
859
46032%
27Gemini 3.5 Flash Lite=
839
32632%
28Hy3 Preview=
822
47628%
29Laguna M.1=
765
46525%
30Nemotron 3 Ultra=
705
43520%
31Claude Haiku 4.5=
705
46520%
32Gemini 3.1 Pro Preview=
696
45419%
33Laguna XS 2.1=
681
46618%
34Mistral Medium 3.5=
641
45115%
35Trinity Large Thinking=
536
4359%