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.
8,083
scored runs
80,557
pairwise comparisons
24
prompts
4,263
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 1Google Gemini
72%±4
Win rate
Elo 1165
Gemini 3.5 Flash-Lite
Tier B
- Won / lost / tied
- 251 / 60 / 245
- prompts
- 8
- comparisons
- 556
- Judge mean
- 83.6
By category
Coding74%
Factual72%
Reasoning75%
Rank 1OpenAI
72%±2
Win rate
Elo 1164
gpt-4.1-mini
Tier C
- Won / lost / tied
- 1077 / 171 / 1206
- prompts
- 24
- comparisons
- 2454
- Judge mean
- 94.8
By category
Coding80%
Creative89%
Factual72%
Multilingual61%
Reasoning77%
Healthcare67%
Rank 1OpenAI
72%±3
Win rate
Elo 1164
gpt-5.6-terra
Tier B
- Won / lost / tied
- 296 / 81 / 251
- prompts
- 9
- comparisons
- 628
- Judge mean
- 86.3
By category
Coding79%
Factual64%
Reasoning75%
Rank 1OpenAI
72%±2
Win rate
Elo 1162
gpt-5.5-2026-04-23
Tier A
- Won / lost / tied
- 998 / 161 / 1149
- prompts
- 20
- comparisons
- 2308
- Judge mean
- 93.9
By category
Coding81%
Creative89%
Factual75%
Multilingual61%
Reasoning69%
Rank 1OpenAI
71%±2
Win rate
Elo 1153
gpt-5.5
Tier C
- Won / lost / tied
- 982 / 198 / 1128
- prompts
- 20
- comparisons
- 2308
- Judge mean
- 92.2
By category
Coding75%
Creative85%
Factual76%
Multilingual58%
Reasoning75%
Rank 1OpenAI
71%±2
Win rate
Elo 1152
gpt-5.1
Tier B
- Won / lost / tied
- 953 / 175 / 1180
- prompts
- 20
- comparisons
- 2308
- Judge mean
- 92.3
By category
Coding78%
Creative78%
Factual65%
Multilingual64%
Reasoning79%
Rank 1OpenAI
70%±4
Win rate
Elo 1146
gpt-6-astra
Tier A
- Won / lost / tied
- 310 / 124 / 194
- prompts
- 9
- comparisons
- 628
- Judge mean
- 87.5
By category
Coding78%
Factual77%
Reasoning43%
Rank 1OpenAI
70%±2
Win rate
Elo 1143
gpt-4.1-2025-04-14
Tier C
- Won / lost / tied
- 964 / 183 / 1307
- prompts
- 24
- comparisons
- 2454
- Judge mean
- 94.8
By category
Coding76%
Creative91%
Factual70%
Multilingual61%
Reasoning69%
Healthcare76%
Rank 1OpenAI
69%±2
Win rate
Elo 1139
gpt-5.2
Tier B
- Won / lost / tied
- 932 / 224 / 1152
- prompts
- 20
- comparisons
- 2308
- Judge mean
- 91.2
By category
Coding64%
Creative88%
Factual70%
Multilingual65%
Reasoning78%
Rank 1OpenAI
69%±2
Win rate
Elo 1138
gpt-5.2-2025-12-11
Tier B
- Won / lost / tied
- 914 / 211 / 1183
- prompts
- 20
- comparisons
- 2308
- Judge mean
- 91.8
By category
Coding77%
Creative79%
Factual53%
Multilingual67%
Reasoning79%
Rank 1OpenAI
68%±2
Win rate
Elo 1134
gpt-5.1-2025-11-13
Tier B
- Won / lost / tied
- 898 / 218 / 1192
- prompts
- 20
- comparisons
- 2308
- Judge mean
- 91.9
By category
Coding77%
Creative75%
Factual63%
Multilingual62%
Reasoning75%
Rank 1OpenAI
68%±2
Win rate
Elo 1134
gpt-5.4-2026-03-05
Tier B
- Won / lost / tied
- 917 / 238 / 1153
- prompts
- 20
- comparisons
- 2308
- Judge mean
- 92.6
By category
Coding75%
Creative90%
Factual54%
Multilingual66%
Reasoning75%
Rank 3OpenAI
68%±2
Win rate
Elo 1130
gpt-4.1-mini-2025-04-14
Tier C
- Won / lost / tied
- 929 / 229 / 1296
- prompts
- 24
- comparisons
- 2454
- Judge mean
- 93.7
By category
Coding72%
Creative85%
Factual60%
Multilingual65%
Reasoning74%
Healthcare62%
Rank 1OpenAI
68%±4
Win rate
Elo 1129
gpt-6.1-sol
Tier A
- Won / lost / tied
- 292 / 132 / 204
- prompts
- 9
- comparisons
- 628
- Judge mean
- 85.8
By category
Coding71%
Factual64%
Reasoning75%
Rank 3Anthropic
68%±2
Win rate
Elo 1129
Claude Opus 4.6
Tier B
- Won / lost / tied
- 971 / 292 / 1136
- prompts
- 24
- comparisons
- 2399
- Judge mean
- 96.3
By category
Coding64%
Creative92%
Factual72%
Multilingual62%
Reasoning68%
Healthcare85%
Rank 7OpenAI
66%±2
Win rate
Elo 1114
gpt-5.4
Tier A
- Won / lost / tied
- 858 / 296 / 1154
- prompts
- 20
- comparisons
- 2308
- Judge mean
- 91.2
By category
Coding73%
Creative84%
Factual51%
Multilingual63%
Reasoning75%
Rank 8OpenAI
66%±2
Win rate
Elo 1112
gpt-4.1
Tier B
- Won / lost / tied
- 903 / 313 / 1238
- prompts
- 24
- comparisons
- 2454
- Judge mean
- 94.1
By category
Coding73%
Creative81%
Factual67%
Multilingual60%
Reasoning62%
Healthcare70%
Rank 3OpenAI
65%±4
Win rate
Elo 1109
gpt-5.6-luna
Tier B
- Won / lost / tied
- 254 / 126 / 248
- prompts
- 9
- comparisons
- 628
- Judge mean
- 84.5
By category
Coding65%
Factual64%
Reasoning75%
Rank 4OpenAI
65%±4
Win rate
Elo 1108
gpt-6-luna
Tier B
- Won / lost / tied
- 304 / 178 / 146
- prompts
- 9
- comparisons
- 628
- Judge mean
- 87.3
By category
Coding68%
Factual87%
Reasoning18%
Rank 10OpenAI
65%±2
Win rate
Elo 1107
gpt-5.4-mini-2026-03-17
Tier A
- Won / lost / tied
- 838 / 318 / 1152
- prompts
- 20
- comparisons
- 2308
- Judge mean
- 89.9
By category
Coding65%
Creative89%
Factual54%
Multilingual60%
Reasoning78%
Rank 10OpenAI
65%±2
Win rate
Elo 1104
gpt-5.4-mini
Tier A
- Won / lost / tied
- 855 / 348 / 1105
- prompts
- 20
- comparisons
- 2308
- Judge mean
- 90.5
By category
Coding71%
Creative86%
Factual55%
Multilingual63%
Reasoning65%
Rank 12Anthropic
64%±2
Win rate
Elo 1103
Claude Opus 4.7
Tier B
- Won / lost / tied
- 900 / 375 / 1124
- prompts
- 24
- comparisons
- 2399
- Judge mean
- 94.4
By category
Coding50%
Creative61%
Factual79%
Multilingual63%
Reasoning74%
Healthcare73%
Rank 15OpenAI
64%±2
Win rate
Elo 1098
o3
Tier C
- Won / lost / tied
- 824 / 364 / 1069
- prompts
- 20
- comparisons
- 2257
- Judge mean
- 85.5
By category
Coding70%
Creative87%
Factual55%
Multilingual61%
Reasoning63%
Rank 15Google Gemini
63%±2
Win rate
Elo 1095
Gemini Flash-Lite Latest
Tier C
- Won / lost / tied
- 906 / 425 / 1123
- prompts
- 24
- comparisons
- 2454
- Judge mean
- 93.6
By category
Coding72%
Creative52%
Factual68%
Multilingual60%
Reasoning60%
Healthcare78%
Rank 15Anthropic
63%±2
Win rate
Elo 1089
Claude Opus 4.5
Tier B
- Won / lost / tied
- 868 / 430 / 1101
- prompts
- 24
- comparisons
- 2399
- Judge mean
- 95.0
By category
Coding57%
Creative61%
Factual69%
Multilingual62%
Reasoning68%
Healthcare83%
Rank 15OpenAI
62%±2
Win rate
Elo 1087
gpt-4o
Tier C
- Won / lost / tied
- 855 / 421 / 1178
- prompts
- 24
- comparisons
- 2454
- Judge mean
- 93.1
By category
Coding56%
Creative69%
Factual61%
Multilingual63%
Reasoning74%
Healthcare60%
Rank 16OpenAI
62%±2
Win rate
Elo 1084
- Won / lost / tied
- 855 / 440 / 1159
- prompts
- 24
- comparisons
- 2454
- Judge mean
- 92.4
By category
Coding60%
Creative77%
Factual56%
Multilingual54%
Reasoning77%
Healthcare83%
Rank 17OpenAI
62%±2
Win rate
Elo 1083
gpt-5-search-api
Tier C
- Won / lost / tied
- 855 / 445 / 1154
- prompts
- 24
- comparisons
- 2454
- Judge mean
- 92.8
By category
Coding64%
Creative83%
Factual51%
Multilingual56%
Reasoning72%
Healthcare82%
Rank 18Anthropic
62%±2
Win rate
Elo 1082
Claude Fable 5
Tier A
- Won / lost / tied
- 790 / 419 / 1074
- prompts
- 20
- comparisons
- 2283
- Judge mean
- 87.5
By category
Coding65%
Creative69%
Factual56%
Multilingual59%
Reasoning70%
Rank 18OpenAI
61%±2
Win rate
Elo 1079
o3-2025-04-16
Tier B
- Won / lost / tied
- 751 / 412 / 1024
- prompts
- 20
- comparisons
- 2187
- Judge mean
- 85.2
By category
Coding64%
Creative86%
Factual51%
Multilingual60%
Reasoning63%
Rank 20OpenAI
61%±2
Win rate
Elo 1074
gpt-4o-2024-08-06
Tier C
- Won / lost / tied
- 829 / 478 / 1147
- prompts
- 24
- comparisons
- 2454
- Judge mean
- 91.0
By category
Coding51%
Creative73%
Factual70%
Multilingual59%
Reasoning64%
Healthcare66%
Rank 21Anthropic
60%±2
Win rate
Elo 1073
Claude Sonnet 4.5
Tier B
- Won / lost / tied
- 818 / 478 / 1103
- prompts
- 24
- comparisons
- 2399
- Judge mean
- 93.4
By category
Coding42%
Creative60%
Factual68%
Multilingual61%
Reasoning78%
Healthcare69%
Rank 19Google Gemini
60%±3
Win rate
Elo 1071
Nano Banana
Tier Barchived
- Won / lost / tied
- 455 / 233 / 696
- prompts
- 23
- comparisons
- 1384
- Judge mean
- 93.5
By category
Coding51%
Creative67%
Factual57%
Multilingual63%
Reasoning74%
Healthcare50%
Rank 21OpenAI
60%±2
Win rate
Elo 1070
gpt-4o-2024-11-20
Tier C
- Won / lost / tied
- 800 / 474 / 1180
- prompts
- 24
- comparisons
- 2454
- Judge mean
- 92.9
By category
Coding61%
Creative52%
Factual61%
Multilingual58%
Reasoning67%
Healthcare78%
Rank 21Z.ai (GLM / Zhipu)
60%±2
Win rate
Elo 1070
GLM-4.7
Tier A
- Won / lost / tied
- 628 / 395 / 653
- prompts
- 14
- comparisons
- 1676
- Judge mean
- 88.9
By category
Coding55%
Creative83%
Factual62%
Reasoning62%
Rank 21Google Gemini
60%±2
Win rate
Elo 1069
Gemini 3.1 Flash Lite
Tier B
- Won / lost / tied
- 735 / 438 / 1135
- prompts
- 20
- comparisons
- 2308
- Judge mean
- 89.9
By category
Coding53%
Creative41%
Factual69%
Multilingual60%
Reasoning69%
Rank 21Anthropic
60%±2
Win rate
Elo 1069
Claude Opus 4.8
Tier A
- Won / lost / tied
- 770 / 481 / 1002
- prompts
- 20
- comparisons
- 2253
- Judge mean
- 92.7
By category
Coding53%
Creative40%
Factual69%
Multilingual62%
Reasoning68%
Rank 21OVH AI Endpoints (GRA)
60%±2
Win rate
Elo 1068
gpt-oss-120b
Tier C
- Won / lost / tied
- 779 / 492 / 1037
- prompts
- 20
- comparisons
- 2308
- Judge mean
- 90.0
By category
Coding62%
Creative84%
Factual58%
Multilingual56%
Reasoning59%
Rank 18OpenAI
59%±4
Win rate
Elo 1067
gpt-5.6-sol
Tier B
- Won / lost / tied
- 239 / 181 / 208
- prompts
- 9
- comparisons
- 628
- Judge mean
- 81.9
By category
Coding55%
Factual60%
Reasoning75%
Rank 21Z.ai (GLM / Zhipu)
59%±3
Win rate
Elo 1065
GLM-5
Tier A
- Won / lost / tied
- 422 / 242 / 575
- prompts
- 14
- comparisons
- 1239
- Judge mean
- 92.7
By category
Coding59%
Creative87%
Factual54%
Reasoning61%
Rank 25Anthropic
59%±2
Win rate
Elo 1063
Claude Sonnet 4.6
Tier A
- Won / lost / tied
- 778 / 500 / 1121
- prompts
- 24
- comparisons
- 2399
- Judge mean
- 93.6
By category
Coding59%
Creative55%
Factual63%
Multilingual53%
Reasoning68%
Healthcare75%
Rank 25Z.ai (GLM / Zhipu)
58%±2
Win rate
Elo 1059
GLM-4.5
Tier A
- Won / lost / tied
- 639 / 456 / 581
- prompts
- 14
- comparisons
- 1676
- Judge mean
- 87.5
By category
Coding57%
Creative85%
Factual67%
Reasoning43%
Rank 25Anthropic
58%±3
Win rate
Elo 1056
Claude Sonnet 5
Tier A
- Won / lost / tied
- 547 / 426 / 529
- prompts
- 19
- comparisons
- 1502
- Judge mean
- 84.0
By category
Coding59%
Creative84%
Factual48%
Multilingual58%
Reasoning65%
Rank 21Google Gemini
58%±4
Win rate
Elo 1053
Gemini 3.7 Flash
Tier B
- Won / lost / tied
- 204 / 174 / 178
- prompts
- 8
- comparisons
- 556
- Judge mean
- 75.7
By category
Coding65%
Factual58%
Reasoning48%
Rank 26Google Gemini
57%±3
Win rate
Elo 1052
Nano Banana 2 (preview)
Tier B
- Won / lost / tied
- 372 / 240 / 472
- prompts
- 18
- comparisons
- 1084
- Judge mean
- 92.6
By category
Coding71%
Creative61%
Factual48%
Multilingual57%
Reasoning64%
Healthcare35%
Rank 30OpenAI
57%±2
Win rate
Elo 1051
gpt-5.4-nano-2026-03-17
Tier A
- Won / lost / tied
- 746 / 558 / 1004
- prompts
- 20
- comparisons
- 2308
- Judge mean
- 82.6
By category
Coding67%
Creative30%
Factual38%
Multilingual66%
Reasoning69%
Rank 30OpenAI
57%±2
Win rate
Elo 1050
gpt-4o-mini
Tier C
- Won / lost / tied
- 762 / 568 / 1124
- prompts
- 24
- comparisons
- 2454
- Judge mean
- 87.0
By category
Coding60%
Creative66%
Factual38%
Multilingual62%
Reasoning70%
Healthcare49%
Rank 33OpenAI
56%±2
Win rate
Elo 1042
gpt-5-mini
Tier C
- Won / lost / tied
- 624 / 497 / 1001
- prompts
- 20
- comparisons
- 2122
- Judge mean
- 82.4
By category
Coding67%
Creative69%
Factual48%
Multilingual52%
Reasoning55%
Rank 29OpenAI
56%±4
Win rate
Elo 1040
gpt-6-sol
Tier A
- Won / lost / tied
- 225 / 211 / 192
- prompts
- 9
- comparisons
- 628
- Judge mean
- 83.1
By category
Coding71%
Factual46%
Reasoning42%
Rank 39OpenAI
55%±2
Win rate
Elo 1036
gpt-4o-mini-2024-07-18
Tier C
- Won / lost / tied
- 748 / 641 / 1065
- prompts
- 24
- comparisons
- 2454
- Judge mean
- 86.6
By category
Coding62%
Creative66%
Factual41%
Multilingual55%
Reasoning65%
Healthcare46%
Rank 43OVH AI Endpoints (GRA)
53%±2
Win rate
Elo 1023
- Won / lost / tied
- 650 / 632 / 1026
- prompts
- 20
- comparisons
- 2308
- Judge mean
- 88.5
By category
Coding47%
Creative74%
Factual50%
Multilingual43%
Reasoning75%
Rank 43OVH AI Endpoints (GRA)
53%±2
Win rate
Elo 1023
Qwen2.5-VL-72B-Instruct
Tier B
- Won / lost / tied
- 691 / 676 / 941
- prompts
- 20
- comparisons
- 2308
- Judge mean
- 88.5
By category
Coding56%
Creative65%
Factual28%
Multilingual55%
Reasoning75%
Rank 44OpenAI
53%±2
Win rate
Elo 1022
gpt-5-mini-2025-08-07
Tier B
- Won / lost / tied
- 603 / 586 / 933
- prompts
- 20
- comparisons
- 2122
- Judge mean
- 80.1
By category
Coding69%
Creative36%
Factual43%
Multilingual47%
Reasoning63%
Rank 47Anthropic
52%±2
Win rate
Elo 1014
Claude Haiku 4.5
Tier A
- Won / lost / tied
- 702 / 736 / 961
- prompts
- 24
- comparisons
- 2399
- Judge mean
- 90.8
By category
Coding61%
Creative42%
Factual50%
Multilingual48%
Reasoning60%
Healthcare30%
Rank 47OpenAI
52%±2
Win rate
Elo 1013
gpt-5.4-nano
Tier C
- Won / lost / tied
- 731 / 774 / 803
- prompts
- 20
- comparisons
- 2308
- Judge mean
- 80.8
By category
Coding57%
Creative33%
Factual24%
Multilingual68%
Reasoning65%
Rank 47Anthropic
52%±2
Win rate
Elo 1011
Claude Opus 5
Tier A
- Won / lost / tied
- 564 / 615 / 676
- prompts
- 20
- comparisons
- 1855
- Judge mean
- 83.2
By category
Coding31%
Creative88%
Factual49%
Multilingual52%
Reasoning56%
Rank 48Google Gemini
51%±3
Win rate
Elo 1004
Gemini Robotics-ER 1.6 Preview
Tier Barchived
- Won / lost / tied
- 383 / 402 / 599
- prompts
- 23
- comparisons
- 1384
- Judge mean
- 79.3
By category
Coding46%
Creative81%
Factual40%
Multilingual56%
Reasoning46%
Healthcare41%
Rank 50OVH AI Endpoints (GRA)
50%±2
Win rate
Elo 1001
- Won / lost / tied
- 654 / 773 / 881
- prompts
- 20
- comparisons
- 2308
- Judge mean
- 87.3
By category
Coding67%
Creative42%
Factual41%
Multilingual45%
Reasoning51%
Rank 53Google Gemini
49%±2
Win rate
Elo 992
Gemini 2.5 Flash-Lite
Tier B
- Won / lost / tied
- 710 / 890 / 854
- prompts
- 24
- comparisons
- 2454
- Judge mean
- 88.9
By category
Coding37%
Creative41%
Factual41%
Multilingual59%
Reasoning63%
Healthcare48%
Rank 60Google Gemini
44%±2
Win rate
Elo 958
Gemini 3 Flash Preview
Tier C
- Won / lost / tied
- 632 / 1027 / 795
- prompts
- 24
- comparisons
- 2454
- Judge mean
- 77.3
By category
Coding32%
Creative23%
Factual51%
Multilingual48%
Reasoning56%
Healthcare25%
Rank 60OpenAI
44%±2
Win rate
Elo 956
gpt-5
Tier C
- Won / lost / tied
- 378 / 651 / 727
- prompts
- 19
- comparisons
- 1756
- Judge mean
- 71.6
By category
Coding58%
Creative8%
Factual42%
Multilingual37%
Reasoning52%
Rank 60OpenAI
42%±2
Win rate
Elo 943
gpt-5-2025-08-07
Tier B
- Won / lost / tied
- 433 / 804 / 713
- prompts
- 20
- comparisons
- 1950
- Judge mean
- 68.2
By category
Coding57%
Creative3%
Factual20%
Multilingual45%
Reasoning55%
Rank 60Google Gemini
38%±4
Win rate
Elo 918
Gemini 3.8 Flash
Tier A
- Won / lost / tied
- 136 / 299 / 121
- prompts
- 8
- comparisons
- 556
- Judge mean
- 63.6
By category
Coding26%
Factual45%
Reasoning48%
Rank 63OVH AI Endpoints (GRA)
35%±2
Win rate
Elo 896
gpt-oss-20b
Tier C
- Won / lost / tied
- 443 / 1126 / 621
- prompts
- 20
- comparisons
- 2190
- Judge mean
- 80.5
By category
Coding48%
Creative42%
Factual17%
Multilingual28%
Reasoning48%
Rank 63Z.ai (GLM / Zhipu)
35%±3
Win rate
Elo 895
GLM-5.1
Tier A
- Won / lost / tied
- 293 / 673 / 333
- prompts
- 14
- comparisons
- 1299
- Judge mean
- 73.4
By category
Coding32%
Creative15%
Factual51%
Reasoning27%
Rank 63Z.ai (GLM / Zhipu)
35%±3
Win rate
Elo 893
GLM-4.5 Air
Tier B
- Won / lost / tied
- 316 / 801 / 449
- prompts
- 14
- comparisons
- 1566
- Judge mean
- 62.9
By category
Coding53%
Creative45%
Factual23%
Reasoning20%
Rank 63Google Gemini
35%±2
Win rate
Elo 889
Gemini Flash Latest
Tier B
- Won / lost / tied
- 543 / 1353 / 558
- prompts
- 24
- comparisons
- 2454
- Judge mean
- 64.1
By category
Coding32%
Creative41%
Factual39%
Multilingual35%
Reasoning31%
Healthcare14%
Rank 63Z.ai (GLM / Zhipu)
33%±3
Win rate
Elo 878
GLM-5 Turbo
Tier B
- Won / lost / tied
- 261 / 767 / 451
- prompts
- 14
- comparisons
- 1479
- Judge mean
- 59.4
By category
Coding44%
Creative32%
Factual38%
Reasoning11%
Rank 63OpenAI
32%±2
Win rate
Elo 870
gpt-5-nano-2025-08-07
Tier B
- Won / lost / tied
- 261 / 817 / 474
- prompts
- 19
- comparisons
- 1552
- Judge mean
- 51.7
By category
Coding27%
Creative8%
Factual51%
Multilingual29%
Reasoning24%
Rank 64Z.ai (GLM / Zhipu)
31%±2
Win rate
Elo 862
GLM-4.6
Tier A
- Won / lost / tied
- 271 / 867 / 413
- prompts
- 14
- comparisons
- 1551
- Judge mean
- 65.2
By category
Coding44%
Creative1%
Factual32%
Reasoning26%
Rank 68OVH AI Endpoints (GRA)
30%±2
Win rate
Elo 852
- Won / lost / tied
- 447 / 1403 / 458
- prompts
- 20
- comparisons
- 2308
- Judge mean
- 73.5
By category
Coding37%
Creative6%
Factual23%
Multilingual34%
Reasoning32%
Rank 68OVH AI Endpoints (GRA)
29%±2
Win rate
Elo 846
Qwen3.5-397B-A17B
Tier A
- Won / lost / tied
- 302 / 1003 / 391
- prompts
- 19
- comparisons
- 1696
- Judge mean
- 49.6
By category
Coding53%
Creative7%
Factual5%
Multilingual35%
Reasoning6%
Rank 68Z.ai (GLM / Zhipu)
29%±3
Win rate
Elo 844
GLM-4.5V (vision)
Tier A
- Won / lost / tied
- 229 / 790 / 349
- prompts
- 14
- comparisons
- 1368
- Judge mean
- 61.9
By category
Coding40%
Creative9%
Factual38%
Reasoning9%
Rank 68Z.ai (GLM / Zhipu)
28%±2
Win rate
Elo 838
GLM-4.6V (vision)
Tier A
- Won / lost / tied
- 207 / 824 / 411
- prompts
- 14
- comparisons
- 1442
- Judge mean
- 54.8
By category
Coding52%
Creative4%
Factual16%
Reasoning19%
Rank 69Z.ai (GLM / Zhipu)
28%±2
Win rate
Elo 838
GLM-5.2
Tier A
- Won / lost / tied
- 255 / 932 / 375
- prompts
- 14
- comparisons
- 1562
- Judge mean
- 55.9
By category
Coding35%
Creative4%
Factual39%
Reasoning8%
Rank 70Google Gemini
27%±2
Win rate
Elo 828
Gemini 3.5 Flash
Tier A
- Won / lost / tied
- 395 / 1468 / 445
- prompts
- 20
- comparisons
- 2308
- Judge mean
- 51.0
By category
Coding8%
Creative53%
Factual46%
Multilingual30%
Reasoning11%
Rank 71OpenAI
26%±2
Win rate
Elo 817
gpt-5-nano
Tier C
- Won / lost / tied
- 176 / 911 / 481
- prompts
- 20
- comparisons
- 1568
- Judge mean
- 50.3
By category
Coding18%
Creative4%
Factual45%
Multilingual25%
Reasoning15%
Rank 77Google Gemini
23%±2
Win rate
Elo 789
Gemini 3.1 Pro Preview
Tier C
- Won / lost / tied
- 384 / 1714 / 356
- prompts
- 24
- comparisons
- 2454
- Judge mean
- 48.1
By category
Coding6%
Creative54%
Factual42%
Multilingual22%
Reasoning9%
Healthcare9%
Rank 77Google Gemini
23%±2
Win rate
Elo 788
- Won / lost / tied
- 210 / 960 / 273
- prompts
- 24
- comparisons
- 1443
- Judge mean
- 44.1
By category
Coding5%
Creative37%
Factual38%
Multilingual22%
Reasoning17%
Healthcare6%
Rank 77OVH AI Endpoints (GRA)
23%±2
Win rate
Elo 787
Qwen3.5-9B
Tier B
- Won / lost / tied
- 194 / 1089 / 398
- prompts
- 19
- comparisons
- 1681
- Judge mean
- 48.1
By category
Coding37%
Creative6%
Factual9%
Multilingual28%
Reasoning5%
Rank 76Google Gemini
23%±3
Win rate
Elo 786
- Won / lost / tied
- 141 / 699 / 244
- prompts
- 18
- comparisons
- 1084
- Judge mean
- 41.4
By category
Coding19%
Creative14%
Factual40%
Multilingual5%
Reasoning4%
Healthcare53%
Rank 79Google Gemini
19%±2
Win rate
Elo 751
Gemini Pro Latest
Tier C
- Won / lost / tied
- 316 / 1810 / 328
- prompts
- 24
- comparisons
- 2454
- Judge mean
- 44.0
By category
Coding5%
Creative27%
Factual40%
Multilingual19%
Reasoning8%
Healthcare5%
Rank 82Google Gemini
18%±2
Win rate
Elo 737
Gemini 2.5 Pro
Tier A
- Won / lost / tied
- 136 / 1106 / 336
- prompts
- 24
- comparisons
- 1578
- Judge mean
- 33.6
By category
Coding16%
Creative3%
Factual27%
Multilingual16%
Reasoning4%
Healthcare53%
Rank 82Google Gemini
16%±2
Win rate
Elo 714
Gemini 2.5 Flash
Tier A
- Won / lost / tied
- 269 / 1908 / 277
- prompts
- 24
- comparisons
- 2454
- Judge mean
- 36.4
By category
Coding4%
Creative23%
Factual21%
Multilingual17%
Reasoning17%
Healthcare4%
Rank 82Google Gemini
15%±3
Win rate
Elo 694
Gemini 3.6 Flash
Tier B
- Won / lost / tied
- 47 / 446 / 63
- prompts
- 8
- comparisons
- 556
- Judge mean
- 44.8
By category
Coding1%
Factual34%
Reasoning5%
85 models ranked from 80,557 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 192
CogView-4
Tier B
- Won / lost / tied
- 0 / 281 / 6
- prompts
- 4
- comparisons
- 287
- Judge mean
- 0.0
By category
Coding1%
Creative<1% · thin