Benchmarks
Intelligence test
Models are ranked by pairwise comparison, not by an absolute score. On every prompt each pair of models is compared head-to-head, and a Bradley-Terry model fits a rating from those comparisons. The win rate below turns that rating into the question it answers: how often this model beats an average model on this board. 50% is average.
7,836
scored runs
105,499
pairwise comparisons
24
prompts
4,586
runs without a score (provider error)
22 of 24 prompts separate models at all. The rest are answered equally well by nearly everyone and carry no ranking signal — that is what an absolute score near 100 was really measuring.
- Coding 4/4
- Creative 2/2
- Factual 3/4
- Multilingual 7/8
- Reasoning 4/4
- Healthcare 2/2
Rank 1OpenAI
74%±2
Win rate
Elo 1179
gpt-5.1
Tier B
- Won / lost / tied
- 983 / 38 / 1334
- prompts
- 20
- comparisons
- 2355
- Judge mean
- 97.2
By category
Coding75%
Creative91%
Factual71%
Multilingual68%
Reasoning79%
Rank 1Z.ai (GLM / Zhipu)
74%±2
Win rate
Elo 1178
GLM-4.7
Tier A
- Won / lost / tied
- 790 / 88 / 861
- prompts
- 14
- comparisons
- 1739
- Judge mean
- 95.8
By category
Coding73%
Creative85%
Factual75%
Reasoning74%
Rank 1OpenAI
72%±2
Win rate
Elo 1162
gpt-5.5-2026-04-23
Tier A
- Won / lost / tied
- 924 / 70 / 1361
- prompts
- 20
- comparisons
- 2355
- Judge mean
- 96.9
By category
Coding76%
Creative90%
Factual68%
Multilingual65%
Reasoning74%
Rank 1OpenAI
71%±2
Win rate
Elo 1152
gpt-4.1-mini
Tier C
- Won / lost / tied
- 1044 / 178 / 1334
- prompts
- 24
- comparisons
- 2556
- Judge mean
- 96.3
By category
Coding75%
Creative89%
Factual71%
Multilingual57%
Reasoning77%
Healthcare75%
Rank 1OpenAI
71%±2
Win rate
Elo 1152
gpt-4.1-2025-04-14
Tier C
- Won / lost / tied
- 982 / 119 / 1455
- prompts
- 24
- comparisons
- 2556
- Judge mean
- 96.6
By category
Coding71%
Creative90%
Factual71%
Multilingual63%
Reasoning73%
Healthcare84%
Rank 1Anthropic
70%±2
Win rate
Elo 1149
Claude Opus 4.6
Tier Barchived
- Won / lost / tied
- 1045 / 199 / 1312
- prompts
- 24
- comparisons
- 2556
- Judge mean
- 97.7
By category
Coding61%
Creative91%
Factual73%
Multilingual67%
Reasoning74%
Healthcare90%
Rank 2OpenAI
70%±2
Win rate
Elo 1146
gpt-5.4-mini
Tier A
- Won / lost / tied
- 931 / 168 / 1256
- prompts
- 20
- comparisons
- 2355
- Judge mean
- 96.5
By category
Coding76%
Creative90%
Factual59%
Multilingual61%
Reasoning79%
Rank 2OpenAI
70%±2
Win rate
Elo 1145
gpt-5.2
Tier B
- Won / lost / tied
- 917 / 161 / 1277
- prompts
- 20
- comparisons
- 2355
- Judge mean
- 96.7
By category
Coding66%
Creative90%
Factual72%
Multilingual64%
Reasoning74%
Rank 3OpenAI
69%±2
Win rate
Elo 1142
gpt-5.5
Tier C
- Won / lost / tied
- 919 / 179 / 1257
- prompts
- 20
- comparisons
- 2355
- Judge mean
- 95.1
By category
Coding68%
Creative80%
Factual74%
Multilingual64%
Reasoning74%
Rank 3OpenAI
69%±2
Win rate
Elo 1139
gpt-4.1
Tier B
- Won / lost / tied
- 957 / 170 / 1429
- prompts
- 24
- comparisons
- 2556
- Judge mean
- 96.2
By category
Coding75%
Creative81%
Factual66%
Multilingual62%
Reasoning73%
Healthcare79%
Rank 3OpenAI
69%±2
Win rate
Elo 1138
gpt-5.4-2026-03-05
Tier B
- Won / lost / tied
- 894 / 176 / 1285
- prompts
- 20
- comparisons
- 2355
- Judge mean
- 96.4
By category
Coding71%
Creative90%
Factual57%
Multilingual67%
Reasoning74%
Rank 3OpenAI
68%±2
Win rate
Elo 1132
gpt-5.2-2025-12-11
Tier B
- Won / lost / tied
- 872 / 186 / 1297
- prompts
- 20
- comparisons
- 2355
- Judge mean
- 95.6
By category
Coding76%
Creative78%
Factual52%
Multilingual67%
Reasoning79%
Rank 4OpenAI
68%±2
Win rate
Elo 1131
gpt-5.1-2025-11-13
Tier B
- Won / lost / tied
- 848 / 170 / 1337
- prompts
- 20
- comparisons
- 2355
- Judge mean
- 95.9
By category
Coding72%
Creative76%
Factual62%
Multilingual67%
Reasoning74%
Rank 4OpenAI
68%±2
Win rate
Elo 1130
gpt-4.1-mini-2025-04-14
Tier C
- Won / lost / tied
- 937 / 206 / 1413
- prompts
- 24
- comparisons
- 2556
- Judge mean
- 95.6
By category
Coding71%
Creative76%
Factual64%
Multilingual66%
Reasoning74%
Healthcare70%
Rank 4OpenAI
67%±2
Win rate
Elo 1123
gpt-4o-2024-05-13
Tier Carchived
- Won / lost / tied
- 904 / 221 / 1431
- prompts
- 24
- comparisons
- 2556
- Judge mean
- 96.0
By category
Coding68%
Creative80%
Factual66%
Multilingual61%
Reasoning73%
Healthcare75%
Rank 4OpenAI
67%±2
Win rate
Elo 1122
gpt-5.4
Tier A
- Won / lost / tied
- 850 / 224 / 1281
- prompts
- 20
- comparisons
- 2355
- Judge mean
- 96.1
By category
Coding65%
Creative89%
Factual57%
Multilingual66%
Reasoning74%
Rank 6Anthropic
67%±2
Win rate
Elo 1120
Claude Opus 4.5
Tier Barchived
- Won / lost / tied
- 950 / 287 / 1319
- prompts
- 24
- comparisons
- 2556
- Judge mean
- 96.8
By category
Coding56%
Creative62%
Factual78%
Multilingual66%
Reasoning74%
Healthcare88%
Rank 7OpenAI
66%±2
Win rate
Elo 1117
- Won / lost / tied
- 960 / 311 / 1285
- prompts
- 24
- comparisons
- 2556
- Judge mean
- 95.0
By category
Coding68%
Creative86%
Factual60%
Multilingual54%
Reasoning78%
Healthcare89%
Rank 7OpenAI
66%±2
Win rate
Elo 1117
gpt-4o
Tier C
- Won / lost / tied
- 923 / 280 / 1353
- prompts
- 24
- comparisons
- 2556
- Judge mean
- 95.5
By category
Coding58%
Creative82%
Factual73%
Multilingual62%
Reasoning73%
Healthcare67%
Rank 9Z.ai (GLM / Zhipu)
65%±2
Win rate
Elo 1110
GLM-4.5
Tier A
- Won / lost / tied
- 721 / 291 / 790
- prompts
- 14
- comparisons
- 1802
- Judge mean
- 90.9
By category
Coding75%
Creative80%
Factual64%
Reasoning52%
Rank 11OpenAI
65%±2
Win rate
Elo 1109
gpt-5.4-mini-2026-03-17
Tier A
- Won / lost / tied
- 819 / 272 / 1264
- prompts
- 20
- comparisons
- 2355
- Judge mean
- 95.2
By category
Coding69%
Creative85%
Factual54%
Multilingual60%
Reasoning74%
Rank 12Anthropic
65%±2
Win rate
Elo 1108
Claude Opus 4.7
Tier Barchived
- Won / lost / tied
- 920 / 335 / 1301
- prompts
- 24
- comparisons
- 2556
- Judge mean
- 96.2
By category
Coding51%
Creative61%
Factual81%
Multilingual63%
Reasoning73%
Healthcare81%
Rank 12Google Gemini
65%±2
Win rate
Elo 1106
Gemini Flash-Lite Latest
Tier C
- Won / lost / tied
- 967 / 392 / 1197
- prompts
- 24
- comparisons
- 2556
- Judge mean
- 95.5
By category
Coding68%
Creative59%
Factual69%
Multilingual64%
Reasoning63%
Healthcare85%
Rank 12Anthropic
64%±3
Win rate
Elo 1102
Claude Sonnet 5
Tier Aarchived
- Won / lost / tied
- 513 / 232 / 639
- prompts
- 15
- comparisons
- 1384
- Judge mean
- 88.7
By category
Coding68%
Creative90%
Factual61%
Multilingual61%
Reasoning53%
Rank 15OpenAI
64%±2
Win rate
Elo 1100
gpt-5-search-api
Tier C
- Won / lost / tied
- 888 / 349 / 1319
- prompts
- 24
- comparisons
- 2556
- Judge mean
- 95.1
By category
Coding62%
Creative82%
Factual59%
Multilingual55%
Reasoning77%
Healthcare87%
Rank 12Z.ai (GLM / Zhipu)
64%±2
Win rate
Elo 1100
GLM-5
Tier A
- Won / lost / tied
- 560 / 217 / 756
- prompts
- 14
- comparisons
- 1533
- Judge mean
- 94.3
By category
Coding62%
Creative89%
Factual64%
Reasoning62%
Rank 18Anthropic
63%±2
Win rate
Elo 1090
Claude Sonnet 4.5
Tier Barchived
- Won / lost / tied
- 889 / 420 / 1247
- prompts
- 24
- comparisons
- 2556
- Judge mean
- 95.2
By category
Coding43%
Creative59%
Factual74%
Multilingual64%
Reasoning77%
Healthcare78%
Rank 19Anthropic
62%±2
Win rate
Elo 1089
Claude Sonnet 4.6
Tier Aarchived
- Won / lost / tied
- 845 / 384 / 1327
- prompts
- 24
- comparisons
- 2556
- Judge mean
- 95.4
By category
Coding55%
Creative55%
Factual66%
Multilingual63%
Reasoning73%
Healthcare82%
Rank 19OpenAI
62%±2
Win rate
Elo 1089
gpt-4-turbo-2024-04-09
Tier Carchived
- Won / lost / tied
- 870 / 409 / 1277
- prompts
- 24
- comparisons
- 2556
- Judge mean
- 94.5
By category
Coding57%
Creative72%
Factual62%
Multilingual60%
Reasoning73%
Healthcare72%
Rank 18Google Gemini
62%±2
Win rate
Elo 1087
Nano Banana
Tier Barchived
- Won / lost / tied
- 615 / 274 / 910
- prompts
- 23
- comparisons
- 1799
- Judge mean
- 93.5
By category
Coding51%
Creative70%
Factual61%
Multilingual65%
Reasoning74%
Healthcare58%
Rank 20OVH AI Endpoints (GRA)
62%±2
Win rate
Elo 1087
gpt-oss-120b
Tier C
- Won / lost / tied
- 793 / 377 / 1185
- prompts
- 20
- comparisons
- 2355
- Judge mean
- 93.0
By category
Coding58%
Creative82%
Factual62%
Multilingual56%
Reasoning70%
Rank 20OpenAI
62%±2
Win rate
Elo 1083
gpt-4o-2024-11-20
Tier C
- Won / lost / tied
- 830 / 404 / 1322
- prompts
- 24
- comparisons
- 2556
- Judge mean
- 95.1
By category
Coding53%
Creative49%
Factual71%
Multilingual61%
Reasoning72%
Healthcare86%
Rank 20OpenAI
62%±2
Win rate
Elo 1082
gpt-4o-mini
Tier C
- Won / lost / tied
- 818 / 403 / 1335
- prompts
- 24
- comparisons
- 2556
- Judge mean
- 90.1
By category
Coding58%
Creative77%
Factual54%
Multilingual66%
Reasoning68%
Healthcare54%
Rank 20OpenAI
61%±2
Win rate
Elo 1081
gpt-5.4-nano-2026-03-17
Tier A
- Won / lost / tied
- 751 / 371 / 1233
- prompts
- 20
- comparisons
- 2355
- Judge mean
- 89.2
By category
Coding71%
Creative31%
Factual49%
Multilingual67%
Reasoning74%
Rank 20Anthropic
61%±2
Win rate
Elo 1080
Claude Fable 5
Tier Aarchived
- Won / lost / tied
- 744 / 364 / 1288
- prompts
- 20
- comparisons
- 2396
- Judge mean
- 89.1
By category
Coding67%
Creative71%
Factual49%
Multilingual62%
Reasoning68%
Rank 21OpenAI
61%±2
Win rate
Elo 1079
gpt-4.1-nano-2025-04-14
Tier Carchived
- Won / lost / tied
- 826 / 426 / 1304
- prompts
- 24
- comparisons
- 2556
- Judge mean
- 92.9
By category
Coding68%
Creative60%
Factual61%
Multilingual56%
Reasoning68%
Healthcare35%
Rank 23OpenAI
61%±2
Win rate
Elo 1078
gpt-4.1-nano
Tier Carchived
- Won / lost / tied
- 825 / 436 / 1295
- prompts
- 24
- comparisons
- 2556
- Judge mean
- 92.0
By category
Coding68%
Creative66%
Factual66%
Multilingual55%
Reasoning64%
Healthcare21%
Rank 22OpenAI
61%±2
Win rate
Elo 1078
o1
Tier Carchived
- Won / lost / tied
- 684 / 342 / 1132
- prompts
- 20
- comparisons
- 2158
- Judge mean
- 89.1
By category
Coding76%
Creative58%
Factual43%
Multilingual63%
Reasoning60%
Rank 24OpenAI
61%±2
Win rate
Elo 1076
gpt-4o-2024-08-06
Tier C
- Won / lost / tied
- 838 / 457 / 1261
- prompts
- 24
- comparisons
- 2556
- Judge mean
- 93.3
By category
Coding54%
Creative76%
Factual68%
Multilingual60%
Reasoning60%
Healthcare72%
Rank 24Google Gemini
60%±3
Win rate
Elo 1070
Nano Banana 2
Tier B
- Won / lost / tied
- 501 / 290 / 608
- prompts
- 18
- comparisons
- 1399
- Judge mean
- 92.6
By category
Coding70%
Creative61%
Factual53%
Multilingual68%
Reasoning65%
Healthcare36%
Rank 25Anthropic
60%±2
Win rate
Elo 1069
Claude Opus 4.8
Tier Aarchived
- Won / lost / tied
- 748 / 439 / 1168
- prompts
- 20
- comparisons
- 2355
- Judge mean
- 95.0
By category
Coding49%
Creative44%
Factual70%
Multilingual61%
Reasoning74%
Rank 25OpenAI
60%±2
Win rate
Elo 1068
gpt-4-turbo
Tier Carchived
- Won / lost / tied
- 818 / 491 / 1247
- prompts
- 24
- comparisons
- 2556
- Judge mean
- 94.2
By category
Coding48%
Creative65%
Factual70%
Multilingual58%
Reasoning65%
Healthcare78%
Rank 26OpenAI
59%±2
Win rate
Elo 1067
o1-2024-12-17
Tier Carchived
- Won / lost / tied
- 629 / 349 / 1180
- prompts
- 20
- comparisons
- 2158
- Judge mean
- 85.1
By category
Coding71%
Creative1%
Factual60%
Multilingual66%
Reasoning60%
Rank 27OpenAI
59%±2
Win rate
Elo 1063
gpt-4o-mini-2024-07-18
Tier C
- Won / lost / tied
- 801 / 511 / 1244
- prompts
- 24
- comparisons
- 2556
- Judge mean
- 89.4
By category
Coding61%
Creative76%
Factual49%
Multilingual61%
Reasoning63%
Healthcare49%
Rank 27Google Gemini
59%±2
Win rate
Elo 1062
Gemini 3.1 Flash Lite
Tier Barchived
- Won / lost / tied
- 687 / 422 / 1246
- prompts
- 20
- comparisons
- 2355
- Judge mean
- 94.2
By category
Coding53%
Creative44%
Factual60%
Multilingual62%
Reasoning74%
Rank 27OpenAI
59%±2
Win rate
Elo 1062
o4-mini-2025-04-16
Tier Barchived
- Won / lost / tied
- 631 / 377 / 1149
- prompts
- 20
- comparisons
- 2157
- Judge mean
- 79.9
By category
Coding60%
Creative66%
Factual45%
Multilingual68%
Reasoning61%
Rank 30OpenAI
58%±2
Win rate
Elo 1058
o4-mini
Tier Carchived
- Won / lost / tied
- 640 / 412 / 1095
- prompts
- 20
- comparisons
- 2147
- Judge mean
- 78.7
By category
Coding61%
Creative77%
Factual50%
Multilingual60%
Reasoning54%
Rank 31OpenAI
58%±2
Win rate
Elo 1055
o3-mini-2025-01-31
Tier Carchived
- Won / lost / tied
- 663 / 443 / 1186
- prompts
- 20
- comparisons
- 2292
- Judge mean
- 89.7
By category
Coding67%
Creative69%
Factual42%
Multilingual61%
Reasoning57%
Rank 37OVH AI Endpoints (GRA)
57%±2
Win rate
Elo 1049
- Won / lost / tied
- 677 / 492 / 1186
- prompts
- 20
- comparisons
- 2355
- Judge mean
- 93.1
By category
Coding45%
Creative63%
Factual63%
Multilingual53%
Reasoning74%
Rank 37OpenAI
57%±2
Win rate
Elo 1048
o3-2025-04-16
Tier Barchived
- Won / lost / tied
- 561 / 397 / 1109
- prompts
- 19
- comparisons
- 2067
- Judge mean
- 79.5
By category
Coding63%
Creative85%
Factual43%
Multilingual53%
Reasoning54%
Rank 40OpenAI
57%±2
Win rate
Elo 1046
o3
Tier C
- Won / lost / tied
- 604 / 441 / 1112
- prompts
- 20
- comparisons
- 2157
- Judge mean
- 80.8
By category
Coding55%
Creative86%
Factual57%
Multilingual49%
Reasoning54%
Rank 47OVH AI Endpoints (GRA)
54%±2
Win rate
Elo 1031
- Won / lost / tied
- 729 / 656 / 970
- prompts
- 20
- comparisons
- 2355
- Judge mean
- 90.6
By category
Coding60%
Creative52%
Factual52%
Multilingual49%
Reasoning62%
Rank 47OVH AI Endpoints (GRA)
54%±2
Win rate
Elo 1029
Qwen2.5-VL-72B-Instruct
Tier B
- Won / lost / tied
- 701 / 639 / 1015
- prompts
- 20
- comparisons
- 2355
- Judge mean
- 92.8
By category
Coding56%
Creative67%
Factual34%
Multilingual55%
Reasoning74%
Rank 47OVH AI Endpoints (GRA)
54%±2
Win rate
Elo 1028
- Won / lost / tied
- 677 / 623 / 1055
- prompts
- 20
- comparisons
- 2355
- Judge mean
- 92.6
By category
Coding62%
Creative49%
Factual50%
Multilingual51%
Reasoning58%
Rank 48Anthropic
54%±2
Win rate
Elo 1028
Claude Haiku 4.5
Tier Aarchived
- Won / lost / tied
- 737 / 682 / 1137
- prompts
- 24
- comparisons
- 2556
- Judge mean
- 92.0
By category
Coding60%
Creative43%
Factual50%
Multilingual57%
Reasoning58%
Healthcare33%
Rank 52OpenAI
52%±2
Win rate
Elo 1013
gpt-5.4-nano
Tier C
- Won / lost / tied
- 686 / 728 / 941
- prompts
- 20
- comparisons
- 2355
- Judge mean
- 87.5
By category
Coding55%
Creative32%
Factual33%
Multilingual66%
Reasoning64%
Rank 52OpenAI
51%±2
Win rate
Elo 1005
o3-mini
Tier Carchived
- Won / lost / tied
- 554 / 630 / 1045
- prompts
- 20
- comparisons
- 2229
- Judge mean
- 83.3
By category
Coding60%
Creative32%
Factual31%
Multilingual59%
Reasoning61%
Rank 52Google Gemini
51%±2
Win rate
Elo 1003
Gemini Robotics-ER 1.6 Preview
Tier Barchived
- Won / lost / tied
- 485 / 536 / 778
- prompts
- 23
- comparisons
- 1799
- Judge mean
- 79.3
By category
Coding44%
Creative83%
Factual43%
Multilingual56%
Reasoning43%
Healthcare43%
Rank 52Google Gemini
50%±2
Win rate
Elo 1003
Gemini 2.5 Flash-Lite
Tier Barchived
- Won / lost / tied
- 736 / 850 / 970
- prompts
- 24
- comparisons
- 2556
- Judge mean
- 91.4
By category
Coding35%
Creative59%
Factual49%
Multilingual59%
Reasoning58%
Healthcare53%
Rank 56Anthropic
49%±2
Win rate
Elo 990
Claude Opus 5
Tier Aarchived
- Won / lost / tied
- 471 / 618 / 771
- prompts
- 17
- comparisons
- 1860
- Judge mean
- 81.4
By category
Coding21%
Creative81%
Factual48%
Multilingual42%
Reasoning61%
Rank 56OpenAI
48%±2
Win rate
Elo 989
gpt-4-0613
Tier Carchived
- Won / lost / tied
- 691 / 897 / 968
- prompts
- 24
- comparisons
- 2556
- Judge mean
- 90.8
By category
Coding50%
Creative45%
Factual38%
Multilingual50%
Reasoning62%
Healthcare51%
Rank 56OpenAI
48%±2
Win rate
Elo 988
gpt-4
Tier Carchived
- Won / lost / tied
- 694 / 905 / 957
- prompts
- 24
- comparisons
- 2556
- Judge mean
- 88.9
By category
Coding49%
Creative34%
Factual34%
Multilingual60%
Reasoning58%
Healthcare43%
Rank 57OpenAI
46%±2
Win rate
Elo 974
gpt-5-mini
Tier C
- Won / lost / tied
- 417 / 646 / 1023
- prompts
- 20
- comparisons
- 2086
- Judge mean
- 71.6
By category
Coding53%
Creative51%
Factual40%
Multilingual39%
Reasoning54%
Rank 60OpenAI
45%±2
Win rate
Elo 968
gpt-3.5-turbo-1106
Tier Carchived
- Won / lost / tied
- 613 / 954 / 989
- prompts
- 24
- comparisons
- 2556
- Judge mean
- 82.7
By category
Coding54%
Creative67%
Factual22%
Multilingual52%
Reasoning48%
Healthcare18%
Rank 60OpenAI
45%±2
Win rate
Elo 963
gpt-3.5-turbo-0125
Tier Carchived
- Won / lost / tied
- 610 / 985 / 961
- prompts
- 24
- comparisons
- 2556
- Judge mean
- 83.4
By category
Coding56%
Creative50%
Factual20%
Multilingual53%
Reasoning51%
Healthcare25%
Rank 63OpenAI
42%±2
Win rate
Elo 946
gpt-5-mini-2025-08-07
Tier Barchived
- Won / lost / tied
- 406 / 805 / 946
- prompts
- 20
- comparisons
- 2157
- Judge mean
- 67.6
By category
Coding52%
Creative36%
Factual40%
Multilingual31%
Reasoning54%
Rank 64OpenAI
42%±2
Win rate
Elo 943
gpt-3.5-turbo
Tier Carchived
- Won / lost / tied
- 607 / 1112 / 837
- prompts
- 24
- comparisons
- 2556
- Judge mean
- 83.1
By category
Coding54%
Creative49%
Factual20%
Multilingual44%
Reasoning49%
Healthcare22%
Rank 64OpenAI
41%±2
Win rate
Elo 940
gpt-3.5-turbo-16k
Tier Carchived
- Won / lost / tied
- 577 / 1101 / 878
- prompts
- 24
- comparisons
- 2556
- Judge mean
- 81.8
By category
Coding57%
Creative31%
Factual23%
Multilingual51%
Reasoning35%
Healthcare20%
Rank 65Google Gemini
41%±2
Win rate
Elo 938
Gemini 3 Flash Preview
Tier C
- Won / lost / tied
- 596 / 1133 / 827
- prompts
- 24
- comparisons
- 2556
- Judge mean
- 78.4
By category
Coding37%
Creative23%
Factual50%
Multilingual43%
Reasoning45%
Healthcare23%
Rank 66Z.ai (GLM / Zhipu)
40%±2
Win rate
Elo 930
GLM-5 Turbo
Tier B
- Won / lost / tied
- 356 / 765 / 681
- prompts
- 14
- comparisons
- 1802
- Judge mean
- 52.1
By category
Coding51%
Creative28%
Factual44%
Reasoning24%
Rank 66Z.ai (GLM / Zhipu)
39%±3
Win rate
Elo 922
GLM-5.1
Tier A
- Won / lost / tied
- 365 / 726 / 442
- prompts
- 14
- comparisons
- 1533
- Judge mean
- 68.8
By category
Coding34%
Creative13%
Factual50%
Reasoning43%
Rank 66Z.ai (GLM / Zhipu)
39%±2
Win rate
Elo 921
GLM-4.5 Air
Tier B
- Won / lost / tied
- 332 / 737 / 599
- prompts
- 14
- comparisons
- 1668
- Judge mean
- 65.4
By category
Coding56%
Creative47%
Factual31%
Reasoning15%
Rank 66Z.ai (GLM / Zhipu)
39%±3
Win rate
Elo 921
GLM-5.2
Tier A
- Won / lost / tied
- 297 / 639 / 515
- prompts
- 14
- comparisons
- 1451
- Judge mean
- 54.9
By category
Coding49%
Creative4%
Factual55%
Reasoning20%
Rank 68OVH AI Endpoints (GRA)
38%±2
Win rate
Elo 912
gpt-oss-20b
Tier C
- Won / lost / tied
- 452 / 1054 / 723
- prompts
- 20
- comparisons
- 2229
- Judge mean
- 79.2
By category
Coding40%
Creative39%
Factual34%
Multilingual27%
Reasoning54%
Rank 70OpenAI
36%±2
Win rate
Elo 904
gpt-5
Tier C
- Won / lost / tied
- 285 / 805 / 749
- prompts
- 19
- comparisons
- 1839
- Judge mean
- 57.4
By category
Coding51%
Creative7%
Factual39%
Multilingual30%
Reasoning37%
Rank 71OVH AI Endpoints (GRA)
35%±2
Win rate
Elo 890
- Won / lost / tied
- 455 / 1195 / 623
- prompts
- 20
- comparisons
- 2273
- Judge mean
- 77.2
By category
Coding47%
Creative9%
Factual29%
Multilingual33%
Reasoning41%
Rank 74Google Gemini
34%±2
Win rate
Elo 885
Gemini Flash Latest
Tier B
- Won / lost / tied
- 522 / 1383 / 651
- prompts
- 24
- comparisons
- 2556
- Judge mean
- 63.1
By category
Coding32%
Creative37%
Factual46%
Multilingual28%
Reasoning31%
Healthcare9%
Rank 74OpenAI
34%±2
Win rate
Elo 881
gpt-5-2025-08-07
Tier Barchived
- Won / lost / tied
- 290 / 964 / 749
- prompts
- 20
- comparisons
- 2003
- Judge mean
- 53.1
By category
Coding35%
Creative3%
Factual24%
Multilingual37%
Reasoning54%
Rank 75Z.ai (GLM / Zhipu)
32%±2
Win rate
Elo 871
GLM-4.5V (vision)
Tier A
- Won / lost / tied
- 257 / 803 / 473
- prompts
- 14
- comparisons
- 1533
- Judge mean
- 52.5
By category
Coding50%
Creative8%
Factual39%
Reasoning6%
Rank 80OVH AI Endpoints (GRA)
27%±2
Win rate
Elo 831
Mistral-7B-Instruct-v0.3
Tier C
- Won / lost / tied
- 395 / 1468 / 492
- prompts
- 20
- comparisons
- 2355
- Judge mean
- 73.3
By category
Coding29%
Creative13%
Factual25%
Multilingual30%
Reasoning30%
Rank 80Z.ai (GLM / Zhipu)
26%±2
Win rate
Elo 817
GLM-4.6
Tier A
- Won / lost / tied
- 185 / 866 / 391
- prompts
- 13
- comparisons
- 1442
- Judge mean
- 48.3
By category
Coding42%
Creative1%
Factual36%
Reasoning5%
Rank 80OpenAI
26%±2
Win rate
Elo 816
gpt-5-nano-2025-08-07
Tier Barchived
- Won / lost / tied
- 197 / 990 / 485
- prompts
- 18
- comparisons
- 1672
- Judge mean
- 36.1
By category
Coding27%
Creative7%
Factual48%
Multilingual18%
Reasoning6%
Rank 80Google Gemini
26%±2
Win rate
Elo 815
Gemini 3.5 Flash
Tier Aarchived
- Won / lost / tied
- 369 / 1522 / 464
- prompts
- 20
- comparisons
- 2355
- Judge mean
- 52.0
By category
Coding8%
Creative54%
Factual39%
Multilingual29%
Reasoning10%
Rank 80OVH AI Endpoints (GRA)
25%±2
Win rate
Elo 812
Qwen3.5-397B-A17B
Tier A
- Won / lost / tied
- 267 / 1235 / 486
- prompts
- 19
- comparisons
- 1988
- Judge mean
- 41.6
By category
Coding50%
Creative6%
Factual6%
Multilingual28%
Reasoning4%
Rank 80Z.ai (GLM / Zhipu)
25%±2
Win rate
Elo 812
GLM-4.6V (vision)
Tier A
- Won / lost / tied
- 186 / 925 / 422
- prompts
- 14
- comparisons
- 1533
- Judge mean
- 45.5
By category
Coding44%
Creative4%
Factual27%
Reasoning5%
Rank 80OVH AI Endpoints (GRA)
25%±2
Win rate
Elo 805
Qwen3-32B
Tier B
- Won / lost / tied
- 418 / 1617 / 320
- prompts
- 20
- comparisons
- 2355
- Judge mean
- 79.0
By category
Coding23%
Creative16%
Factual23%
Multilingual26%
Reasoning27%
Rank 80OpenAI
24%±2
Win rate
Elo 797
gpt-5-nano
Tier C
- Won / lost / tied
- 208 / 1154 / 485
- prompts
- 20
- comparisons
- 1847
- Judge mean
- 36.6
By category
Coding18%
Creative4%
Factual50%
Multilingual18%
Reasoning7%
Rank 80Google Gemini
23%±2
Win rate
Elo 793
Nano Banana Pro
Tier A
- Won / lost / tied
- 192 / 915 / 292
- prompts
- 18
- comparisons
- 1399
- Judge mean
- 41.4
By category
Coding19%
Creative14%
Factual44%
Multilingual4%
Reasoning3%
Healthcare54%
Rank 81Google Gemini
22%±2
Win rate
Elo 784
Gemini 3.1 Pro Preview
Tier C
- Won / lost / tied
- 381 / 1783 / 392
- prompts
- 24
- comparisons
- 2556
- Judge mean
- 48.3
By category
Coding8%
Creative51%
Factual45%
Multilingual17%
Reasoning7%
Healthcare5%
Rank 81Google Gemini
22%±2
Win rate
Elo 780
- Won / lost / tied
- 250 / 1275 / 353
- prompts
- 24
- comparisons
- 1878
- Judge mean
- 44.1
By category
Coding4%
Creative39%
Factual42%
Multilingual19%
Reasoning13%
Healthcare3%
Rank 86Google Gemini
21%±2
Win rate
Elo 772
Gemini Pro Latest
Tier C
- Won / lost / tied
- 356 / 1818 / 382
- prompts
- 24
- comparisons
- 2556
- Judge mean
- 44.5
By category
Coding5%
Creative39%
Factual45%
Multilingual19%
Reasoning8%
Healthcare2%
Rank 87OVH AI Endpoints (GRA)
20%±2
Win rate
Elo 761
Qwen3.5-9B
Tier B
- Won / lost / tied
- 187 / 1348 / 453
- prompts
- 19
- comparisons
- 1988
- Judge mean
- 42.3
By category
Coding33%
Creative5%
Factual11%
Multilingual24%
Reasoning3%
Rank 92Google Gemini
17%±2
Win rate
Elo 726
Gemini 2.5 Pro
Tier Aarchived
- Won / lost / tied
- 182 / 1437 / 345
- prompts
- 24
- comparisons
- 1964
- Judge mean
- 30.7
By category
Coding17%
Creative3%
Factual34%
Multilingual8%
Reasoning3%
Healthcare54%
Rank 93Google Gemini
15%±1
Win rate
Elo 704
Gemini 2.5 Flash
Tier Aarchived
- Won / lost / tied
- 284 / 2026 / 246
- prompts
- 24
- comparisons
- 2556
- Judge mean
- 38.6
By category
Coding4%
Creative26%
Factual28%
Multilingual12%
Reasoning10%
Healthcare2%
94 models ranked from 105,499 pairwise comparisons across 24 prompts.
Provisional
Too little evidence to rank: needs at least 5 prompts and 30 comparisons.
ProvisionalZ.ai (GLM / Zhipu)
1%±1
Win rate
Elo 155
CogView-4
Tier B
- Won / lost / tied
- 0 / 361 / 6
- prompts
- 4
- comparisons
- 367
- Judge mean
- 0.0
By category
Coding1%
Creative<1% · thin