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.
8,879
scored runs
121,133
pairwise comparisons
24
prompts
4,760
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
72%±2
Win rate
Elo 1165
gpt-4.1-mini
Tier C
- Won / lost / tied
- 1308 / 197 / 1485
- prompts
- 24
- comparisons
- 2990
- Judge mean
- 96.0
By category
Coding78%
Creative91%
Factual74%
Multilingual60%
Reasoning76%
Healthcare75%
Rank 1OpenAI
71%±2
Win rate
Elo 1156
gpt-5.1
Tier B
- Won / lost / tied
- 1147 / 174 / 1468
- prompts
- 20
- comparisons
- 2789
- Judge mean
- 95.1
By category
Coding76%
Creative79%
Factual71%
Multilingual63%
Reasoning78%
Rank 1OpenAI
71%±2
Win rate
Elo 1156
gpt-5.5-2026-04-23
Tier A
- Won / lost / tied
- 1171 / 199 / 1419
- prompts
- 20
- comparisons
- 2789
- Judge mean
- 95.8
By category
Coding78%
Creative89%
Factual76%
Multilingual61%
Reasoning66%
Rank 1OpenAI
70%±2
Win rate
Elo 1148
gpt-4.1-2025-04-14
Tier C
- Won / lost / tied
- 1189 / 201 / 1600
- prompts
- 24
- comparisons
- 2990
- Judge mean
- 96.0
By category
Coding74%
Creative92%
Factual71%
Multilingual64%
Reasoning66%
Healthcare84%
Rank 1OpenAI
70%±2
Win rate
Elo 1146
gpt-5.4
Tier A
- Won / lost / tied
- 1115 / 208 / 1466
- prompts
- 20
- comparisons
- 2789
- Judge mean
- 95.0
By category
Coding72%
Creative90%
Factual63%
Multilingual67%
Reasoning74%
Rank 1OpenAI
69%±2
Win rate
Elo 1141
gpt-5.5
Tier C
- Won / lost / tied
- 1122 / 252 / 1415
- prompts
- 20
- comparisons
- 2789
- Judge mean
- 94.0
By category
Coding74%
Creative81%
Factual71%
Multilingual61%
Reasoning74%
Rank 1OpenAI
69%±2
Win rate
Elo 1138
gpt-4.1-mini-2025-04-14
Tier C
- Won / lost / tied
- 1147 / 228 / 1615
- prompts
- 24
- comparisons
- 2990
- Judge mean
- 95.1
By category
Coding71%
Creative81%
Factual66%
Multilingual67%
Reasoning73%
Healthcare70%
Rank 2OpenAI
69%±2
Win rate
Elo 1135
gpt-5.2-2025-12-11
Tier B
- Won / lost / tied
- 1081 / 249 / 1459
- prompts
- 20
- comparisons
- 2789
- Judge mean
- 94.1
By category
Coding76%
Creative78%
Factual57%
Multilingual64%
Reasoning78%
Rank 2OpenAI
68%±2
Win rate
Elo 1135
gpt-5.2
Tier B
- Won / lost / tied
- 1110 / 282 / 1397
- prompts
- 20
- comparisons
- 2789
- Judge mean
- 93.6
By category
Coding65%
Creative91%
Factual67%
Multilingual62%
Reasoning77%
Rank 2Z.ai (GLM / Zhipu)
68%±2
Win rate
Elo 1130
GLM-4.7
Tier A
- Won / lost / tied
- 855 / 268 / 904
- prompts
- 14
- comparisons
- 2027
- Judge mean
- 91.3
By category
Coding62%
Creative84%
Factual70%
Reasoning73%
Rank 4OpenAI
67%±2
Win rate
Elo 1127
gpt-5.4-2026-03-05
Tier B
- Won / lost / tied
- 1071 / 297 / 1421
- prompts
- 20
- comparisons
- 2789
- Judge mean
- 94.7
By category
Coding74%
Creative89%
Factual54%
Multilingual63%
Reasoning73%
Rank 4Anthropic
67%±2
Win rate
Elo 1126
Claude Opus 4.6
Tier B
- Won / lost / tied
- 1183 / 354 / 1453
- prompts
- 24
- comparisons
- 2990
- Judge mean
- 96.9
By category
Coding61%
Creative92%
Factual73%
Multilingual62%
Reasoning66%
Healthcare90%
Rank 4OpenAI
67%±2
Win rate
Elo 1126
gpt-4.1
Tier B
- Won / lost / tied
- 1140 / 311 / 1539
- prompts
- 24
- comparisons
- 2990
- Judge mean
- 95.3
By category
Coding69%
Creative83%
Factual69%
Multilingual64%
Reasoning66%
Healthcare79%
Rank 4Z.ai (GLM / Zhipu)
67%±2
Win rate
Elo 1121
GLM-4.5
Tier A
- Won / lost / tied
- 865 / 325 / 837
- prompts
- 14
- comparisons
- 2027
- Judge mean
- 89.9
By category
Coding74%
Creative85%
Factual66%
Reasoning56%
Rank 5OpenAI
67%±2
Win rate
Elo 1120
gpt-5.4-mini
Tier A
- Won / lost / tied
- 1065 / 339 / 1385
- prompts
- 20
- comparisons
- 2789
- Judge mean
- 93.5
By category
Coding72%
Creative91%
Factual61%
Multilingual63%
Reasoning62%
Rank 6OpenAI
66%±2
Win rate
Elo 1118
gpt-5.4-mini-2026-03-17
Tier A
- Won / lost / tied
- 1036 / 328 / 1425
- prompts
- 20
- comparisons
- 2789
- Judge mean
- 92.5
By category
Coding64%
Creative88%
Factual59%
Multilingual62%
Reasoning77%
Rank 6OpenAI
66%±2
Win rate
Elo 1115
gpt-5.1-2025-11-13
Tier B
- Won / lost / tied
- 1010 / 320 / 1459
- prompts
- 20
- comparisons
- 2789
- Judge mean
- 94.0
By category
Coding76%
Creative73%
Factual59%
Multilingual58%
Reasoning74%
Rank 10OpenAI
64%±2
Win rate
Elo 1102
gpt-4o
Tier C
- Won / lost / tied
- 1058 / 412 / 1520
- prompts
- 24
- comparisons
- 2990
- Judge mean
- 94.7
By category
Coding57%
Creative72%
Factual67%
Multilingual64%
Reasoning73%
Healthcare67%
Rank 14Anthropic
64%±2
Win rate
Elo 1098
Claude Opus 4.7
Tier B
- Won / lost / tied
- 1080 / 463 / 1447
- prompts
- 24
- comparisons
- 2990
- Judge mean
- 94.9
By category
Coding47%
Creative59%
Factual79%
Multilingual63%
Reasoning73%
Healthcare81%
Rank 14OpenAI
64%±2
Win rate
Elo 1098
- Won / lost / tied
- 1089 / 472 / 1429
- prompts
- 24
- comparisons
- 2990
- Judge mean
- 94.0
By category
Coding62%
Creative81%
Factual62%
Multilingual54%
Reasoning76%
Healthcare89%
Rank 14OpenAI
64%±2
Win rate
Elo 1098
gpt-4o-2024-05-13
Tier C
- Won / lost / tied
- 1033 / 421 / 1536
- prompts
- 24
- comparisons
- 2990
- Judge mean
- 95.2
By category
Coding61%
Creative82%
Factual61%
Multilingual63%
Reasoning66%
Healthcare75%
Rank 14Google Gemini
64%±2
Win rate
Elo 1097
Gemini Flash-Lite Latest
Tier C
- Won / lost / tied
- 1102 / 496 / 1392
- prompts
- 24
- comparisons
- 2990
- Judge mean
- 94.8
By category
Coding67%
Creative44%
Factual71%
Multilingual61%
Reasoning64%
Healthcare85%
Rank 14OpenAI
64%±2
Win rate
Elo 1097
gpt-5-search-api
Tier C
- Won / lost / tied
- 1077 / 474 / 1439
- prompts
- 24
- comparisons
- 2990
- Judge mean
- 94.3
By category
Coding61%
Creative85%
Factual61%
Multilingual58%
Reasoning70%
Healthcare87%
Rank 14OpenAI
63%±2
Win rate
Elo 1095
o1
Tier C
- Won / lost / tied
- 931 / 404 / 1321
- prompts
- 20
- comparisons
- 2656
- Judge mean
- 90.8
By category
Coding79%
Creative55%
Factual48%
Multilingual64%
Reasoning62%
Rank 15Anthropic
63%±2
Win rate
Elo 1092
Claude Fable 5
Tier A
- Won / lost / tied
- 947 / 416 / 1467
- prompts
- 20
- comparisons
- 2830
- Judge mean
- 89.2
By category
Coding64%
Creative66%
Factual59%
Multilingual64%
Reasoning68%
Rank 17Anthropic
62%±2
Win rate
Elo 1088
Claude Opus 4.5
Tier B
- Won / lost / tied
- 1038 / 499 / 1453
- prompts
- 24
- comparisons
- 2990
- Judge mean
- 95.8
By category
Coding54%
Creative58%
Factual71%
Multilingual62%
Reasoning66%
Healthcare88%
Rank 16Z.ai (GLM / Zhipu)
62%±3
Win rate
Elo 1086
GLM-5
Tier A
- Won / lost / tied
- 539 / 251 / 731
- prompts
- 14
- comparisons
- 1521
- Judge mean
- 94.6
By category
Coding57%
Creative89%
Factual63%
Reasoning61%
Rank 18OpenAI
62%±2
Win rate
Elo 1084
o3-2025-04-16
Tier B
- Won / lost / tied
- 846 / 417 / 1304
- prompts
- 19
- comparisons
- 2567
- Judge mean
- 83.3
By category
Coding69%
Creative83%
Factual48%
Multilingual61%
Reasoning58%
Rank 18Google Gemini
62%±2
Win rate
Elo 1083
Nano Banana
Tier B
- Won / lost / tied
- 600 / 274 / 907
- prompts
- 23
- comparisons
- 1781
- Judge mean
- 93.5
By category
Coding50%
Creative69%
Factual61%
Multilingual64%
Reasoning73%
Healthcare58%
Rank 18OpenAI
62%±2
Win rate
Elo 1082
o1-2024-12-17
Tier C
- Won / lost / tied
- 840 / 415 / 1331
- prompts
- 20
- comparisons
- 2586
- Judge mean
- 87.4
By category
Coding75%
Creative6%
Factual60%
Multilingual67%
Reasoning63%
Rank 18OpenAI
61%±2
Win rate
Elo 1081
o3
Tier C
- Won / lost / tied
- 846 / 435 / 1230
- prompts
- 20
- comparisons
- 2511
- Judge mean
- 84.2
By category
Coding64%
Creative88%
Factual57%
Multilingual58%
Reasoning58%
Rank 18OpenAI
61%±2
Win rate
Elo 1081
gpt-4o-2024-11-20
Tier C
- Won / lost / tied
- 973 / 493 / 1524
- prompts
- 24
- comparisons
- 2990
- Judge mean
- 94.3
By category
Coding53%
Creative49%
Factual71%
Multilingual59%
Reasoning71%
Healthcare86%
Rank 18Google Gemini
61%±2
Win rate
Elo 1079
Gemini 3.1 Flash Lite
Tier B
- Won / lost / tied
- 884 / 451 / 1454
- prompts
- 20
- comparisons
- 2789
- Judge mean
- 92.8
By category
Coding55%
Creative41%
Factual70%
Multilingual60%
Reasoning74%
Rank 18OpenAI
61%±2
Win rate
Elo 1075
o4-mini
Tier C
- Won / lost / tied
- 892 / 502 / 1262
- prompts
- 20
- comparisons
- 2656
- Judge mean
- 81.8
By category
Coding72%
Creative78%
Factual54%
Multilingual57%
Reasoning52%
Rank 18OpenAI
61%±2
Win rate
Elo 1075
o4-mini-2025-04-16
Tier B
- Won / lost / tied
- 819 / 440 / 1322
- prompts
- 20
- comparisons
- 2581
- Judge mean
- 82.6
By category
Coding58%
Creative68%
Factual50%
Multilingual69%
Reasoning63%
Rank 19OpenAI
60%±2
Win rate
Elo 1073
gpt-4o-2024-08-06
Tier C
- Won / lost / tied
- 999 / 579 / 1412
- prompts
- 24
- comparisons
- 2990
- Judge mean
- 92.1
By category
Coding46%
Creative72%
Factual71%
Multilingual62%
Reasoning61%
Healthcare72%
Rank 19Anthropic
60%±2
Win rate
Elo 1073
Claude Sonnet 4.5
Tier B
- Won / lost / tied
- 990 / 572 / 1428
- prompts
- 24
- comparisons
- 2990
- Judge mean
- 94.0
By category
Coding40%
Creative56%
Factual70%
Multilingual61%
Reasoning76%
Healthcare78%
Rank 22OpenAI
60%±2
Win rate
Elo 1071
gpt-4o-mini
Tier C
- Won / lost / tied
- 970 / 569 / 1451
- prompts
- 24
- comparisons
- 2990
- Judge mean
- 89.0
By category
Coding64%
Creative71%
Factual46%
Multilingual62%
Reasoning69%
Healthcare54%
Rank 24Anthropic
59%±2
Win rate
Elo 1066
Claude Sonnet 5
Tier A
- Won / lost / tied
- 636 / 434 / 759
- prompts
- 19
- comparisons
- 1829
- Judge mean
- 86.1
By category
Coding53%
Creative88%
Factual52%
Multilingual62%
Reasoning60%
Rank 25OVH AI Endpoints (GRA)
59%±2
Win rate
Elo 1066
gpt-oss-120b
Tier C
- Won / lost / tied
- 911 / 575 / 1303
- prompts
- 20
- comparisons
- 2789
- Judge mean
- 90.9
By category
Coding59%
Creative84%
Factual61%
Multilingual52%
Reasoning62%
Rank 21Google Gemini
59%±3
Win rate
Elo 1066
Nano Banana 2
Tier B
- Won / lost / tied
- 489 / 290 / 607
- prompts
- 18
- comparisons
- 1386
- Judge mean
- 92.6
By category
Coding70%
Creative60%
Factual52%
Multilingual66%
Reasoning64%
Healthcare36%
Rank 25Anthropic
59%±2
Win rate
Elo 1065
Claude Opus 4.8
Tier A
- Won / lost / tied
- 909 / 583 / 1297
- prompts
- 20
- comparisons
- 2789
- Judge mean
- 93.9
By category
Coding48%
Creative43%
Factual71%
Multilingual62%
Reasoning66%
Rank 25OpenAI
59%±2
Win rate
Elo 1064
o3-mini-2025-01-31
Tier C
- Won / lost / tied
- 848 / 531 / 1344
- prompts
- 20
- comparisons
- 2723
- Judge mean
- 89.8
By category
Coding67%
Creative71%
Factual41%
Multilingual63%
Reasoning59%
Rank 26OpenAI
59%±2
Win rate
Elo 1064
gpt-4-turbo-2024-04-09
Tier C
- Won / lost / tied
- 984 / 634 / 1372
- prompts
- 24
- comparisons
- 2990
- Judge mean
- 93.4
By category
Coding56%
Creative76%
Factual56%
Multilingual52%
Reasoning73%
Healthcare72%
Rank 26OpenAI
59%±2
Win rate
Elo 1064
gpt-5.4-nano-2026-03-17
Tier A
- Won / lost / tied
- 909 / 592 / 1288
- prompts
- 20
- comparisons
- 2789
- Judge mean
- 86.2
By category
Coding65%
Creative30%
Factual43%
Multilingual66%
Reasoning74%
Rank 27Anthropic
59%±2
Win rate
Elo 1060
Claude Sonnet 4.6
Tier A
- Won / lost / tied
- 947 / 631 / 1412
- prompts
- 24
- comparisons
- 2990
- Judge mean
- 94.3
By category
Coding56%
Creative53%
Factual64%
Multilingual54%
Reasoning66%
Healthcare82%
Rank 27OpenAI
58%±2
Win rate
Elo 1059
gpt-4-turbo
Tier C
- Won / lost / tied
- 927 / 616 / 1447
- prompts
- 24
- comparisons
- 2990
- Judge mean
- 93.3
By category
Coding48%
Creative70%
Factual69%
Multilingual57%
Reasoning59%
Healthcare78%
Rank 27OpenAI
58%±2
Win rate
Elo 1056
gpt-4.1-nano-2025-04-14
Tier C
- Won / lost / tied
- 923 / 636 / 1431
- prompts
- 24
- comparisons
- 2990
- Judge mean
- 91.9
By category
Coding61%
Creative66%
Factual58%
Multilingual51%
Reasoning69%
Healthcare35%
Rank 38OpenAI
57%±2
Win rate
Elo 1046
gpt-4.1-nano
Tier C
- Won / lost / tied
- 887 / 685 / 1418
- prompts
- 24
- comparisons
- 2990
- Judge mean
- 91.1
By category
Coding60%
Creative59%
Factual63%
Multilingual46%
Reasoning68%
Healthcare21%
Rank 39OpenAI
56%±2
Win rate
Elo 1040
gpt-4o-mini-2024-07-18
Tier C
- Won / lost / tied
- 907 / 751 / 1332
- prompts
- 24
- comparisons
- 2990
- Judge mean
- 88.2
By category
Coding60%
Creative70%
Factual43%
Multilingual55%
Reasoning63%
Healthcare49%
Rank 40OVH AI Endpoints (GRA)
55%±2
Win rate
Elo 1038
- Won / lost / tied
- 792 / 664 / 1333
- prompts
- 20
- comparisons
- 2789
- Judge mean
- 91.5
By category
Coding43%
Creative69%
Factual59%
Multilingual50%
Reasoning73%
Rank 49Anthropic
54%±2
Win rate
Elo 1028
Claude Haiku 4.5
Tier A
- Won / lost / tied
- 881 / 820 / 1289
- prompts
- 24
- comparisons
- 2990
- Judge mean
- 91.6
By category
Coding58%
Creative54%
Factual53%
Multilingual52%
Reasoning60%
Healthcare33%
Rank 49OVH AI Endpoints (GRA)
54%±2
Win rate
Elo 1025
- Won / lost / tied
- 845 / 814 / 1130
- prompts
- 20
- comparisons
- 2789
- Judge mean
- 89.1
By category
Coding55%
Creative36%
Factual56%
Multilingual50%
Reasoning64%
Rank 49OpenAI
53%±2
Win rate
Elo 1021
gpt-5.4-nano
Tier C
- Won / lost / tied
- 889 / 889 / 1011
- prompts
- 20
- comparisons
- 2789
- Judge mean
- 84.7
By category
Coding54%
Creative31%
Factual31%
Multilingual70%
Reasoning61%
Rank 51OVH AI Endpoints (GRA)
52%±2
Win rate
Elo 1013
- Won / lost / tied
- 804 / 865 / 1120
- prompts
- 20
- comparisons
- 2789
- Judge mean
- 90.6
By category
Coding66%
Creative40%
Factual45%
Multilingual46%
Reasoning57%
Rank 51OVH AI Endpoints (GRA)
52%±2
Win rate
Elo 1011
Qwen2.5-VL-72B-Instruct
Tier B
- Won / lost / tied
- 784 / 855 / 1150
- prompts
- 20
- comparisons
- 2789
- Judge mean
- 91.3
By category
Coding50%
Creative64%
Factual32%
Multilingual51%
Reasoning74%
Rank 52Anthropic
51%±2
Win rate
Elo 1007
Claude Opus 5
Tier A
- Won / lost / tied
- 656 / 749 / 891
- prompts
- 20
- comparisons
- 2296
- Judge mean
- 82.4
By category
Coding26%
Creative85%
Factual52%
Multilingual51%
Reasoning55%
Rank 52OpenAI
51%±2
Win rate
Elo 1005
gpt-5-mini
Tier C
- Won / lost / tied
- 560 / 641 / 1171
- prompts
- 20
- comparisons
- 2372
- Judge mean
- 75.6
By category
Coding59%
Creative50%
Factual40%
Multilingual47%
Reasoning58%
Rank 52OpenAI
50%±2
Win rate
Elo 1002
o3-mini
Tier C
- Won / lost / tied
- 670 / 797 / 1186
- prompts
- 20
- comparisons
- 2653
- Judge mean
- 85.1
By category
Coding56%
Creative29%
Factual31%
Multilingual54%
Reasoning65%
Rank 52Google Gemini
50%±2
Win rate
Elo 1001
Gemini Robotics-ER 1.6 Preview
Tier Barchived
- Won / lost / tied
- 473 / 532 / 776
- prompts
- 23
- comparisons
- 1781
- Judge mean
- 79.3
By category
Coding44%
Creative82%
Factual43%
Multilingual55%
Reasoning43%
Healthcare43%
Rank 54Google Gemini
49%±2
Win rate
Elo 995
Gemini 2.5 Flash-Lite
Tier B
- Won / lost / tied
- 827 / 1031 / 1132
- prompts
- 24
- comparisons
- 2990
- Judge mean
- 90.5
By category
Coding35%
Creative50%
Factual44%
Multilingual58%
Reasoning60%
Healthcare53%
Rank 57OpenAI
47%±2
Win rate
Elo 981
gpt-5-mini-2025-08-07
Tier B
- Won / lost / tied
- 565 / 810 / 1069
- prompts
- 20
- comparisons
- 2444
- Judge mean
- 73.0
By category
Coding64%
Creative36%
Factual39%
Multilingual35%
Reasoning58%
Rank 60OpenAI
46%±2
Win rate
Elo 974
gpt-4
Tier C
- Won / lost / tied
- 793 / 1157 / 1040
- prompts
- 24
- comparisons
- 2990
- Judge mean
- 87.7
By category
Coding45%
Creative32%
Factual40%
Multilingual48%
Reasoning59%
Healthcare43%
Rank 60OpenAI
46%±2
Win rate
Elo 974
gpt-3.5-turbo-1106
Tier C
- Won / lost / tied
- 746 / 1112 / 1132
- prompts
- 24
- comparisons
- 2990
- Judge mean
- 81.8
By category
Coding55%
Creative54%
Factual23%
Multilingual54%
Reasoning52%
Healthcare18%
Rank 61OpenAI
46%±2
Win rate
Elo 969
gpt-4-0613
Tier Carchived
- Won / lost / tied
- 763 / 1166 / 1061
- prompts
- 24
- comparisons
- 2990
- Judge mean
- 89.3
By category
Coding44%
Creative52%
Factual37%
Multilingual46%
Reasoning54%
Healthcare51%
Rank 63OpenAI
42%±2
Win rate
Elo 947
gpt-5
Tier C
- Won / lost / tied
- 414 / 803 / 916
- prompts
- 19
- comparisons
- 2133
- Judge mean
- 62.8
By category
Coding56%
Creative7%
Factual38%
Multilingual41%
Reasoning45%
Rank 65Google Gemini
42%±2
Win rate
Elo 947
Gemini 3 Flash Preview
Tier C
- Won / lost / tied
- 709 / 1285 / 996
- prompts
- 24
- comparisons
- 2990
- Judge mean
- 77.3
By category
Coding33%
Creative20%
Factual52%
Multilingual46%
Reasoning48%
Healthcare23%
Rank 65OpenAI
42%±2
Win rate
Elo 946
gpt-3.5-turbo-0125
Tier C
- Won / lost / tied
- 692 / 1272 / 1026
- prompts
- 24
- comparisons
- 2990
- Judge mean
- 82.0
By category
Coding57%
Creative45%
Factual18%
Multilingual48%
Reasoning42%
Healthcare25%
Rank 66OpenAI
42%±2
Win rate
Elo 942
gpt-3.5-turbo
Tier C
- Won / lost / tied
- 716 / 1330 / 944
- prompts
- 24
- comparisons
- 2990
- Judge mean
- 82.6
By category
Coding49%
Creative53%
Factual18%
Multilingual44%
Reasoning54%
Healthcare22%
Rank 66OpenAI
41%±2
Win rate
Elo 940
gpt-5-2025-08-07
Tier B
- Won / lost / tied
- 483 / 967 / 922
- prompts
- 20
- comparisons
- 2372
- Judge mean
- 59.9
By category
Coding57%
Creative3%
Factual23%
Multilingual40%
Reasoning58%
Rank 66OpenAI
41%±2
Win rate
Elo 937
gpt-3.5-turbo-16k
Tier C
- Won / lost / tied
- 698 / 1349 / 943
- prompts
- 24
- comparisons
- 2990
- Judge mean
- 81.2
By category
Coding62%
Creative29%
Factual21%
Multilingual46%
Reasoning34%
Healthcare20%
Rank 66OVH AI Endpoints (GRA)
39%±2
Win rate
Elo 920
gpt-oss-20b
Tier C
- Won / lost / tied
- 577 / 1261 / 814
- prompts
- 20
- comparisons
- 2652
- Judge mean
- 80.3
By category
Coding50%
Creative45%
Factual31%
Multilingual27%
Reasoning48%
Rank 69Z.ai (GLM / Zhipu)
38%±2
Win rate
Elo 913
GLM-5 Turbo
Tier B
- Won / lost / tied
- 356 / 838 / 607
- prompts
- 14
- comparisons
- 1801
- Judge mean
- 54.5
By category
Coding43%
Creative31%
Factual44%
Reasoning25%
Rank 72Z.ai (GLM / Zhipu)
37%±2
Win rate
Elo 905
GLM-4.5 Air
Tier B
- Won / lost / tied
- 357 / 873 / 579
- prompts
- 14
- comparisons
- 1809
- Judge mean
- 65.4
By category
Coding50%
Creative46%
Factual31%
Reasoning15%
Rank 72Z.ai (GLM / Zhipu)
36%±2
Win rate
Elo 903
GLM-5.1
Tier A
- Won / lost / tied
- 366 / 823 / 411
- prompts
- 14
- comparisons
- 1600
- Judge mean
- 69.1
By category
Coding32%
Creative13%
Factual50%
Reasoning34%
Rank 72Z.ai (GLM / Zhipu)
36%±2
Win rate
Elo 899
GLM-5.2
Tier A
- Won / lost / tied
- 358 / 901 / 540
- prompts
- 14
- comparisons
- 1799
- Judge mean
- 52.3
By category
Coding50%
Creative4%
Factual42%
Reasoning17%
Rank 73Google Gemini
35%±2
Win rate
Elo 892
Gemini Flash Latest
Tier B
- Won / lost / tied
- 621 / 1600 / 769
- prompts
- 24
- comparisons
- 2990
- Judge mean
- 63.0
By category
Coding29%
Creative31%
Factual49%
Multilingual35%
Reasoning27%
Healthcare9%
Rank 73Z.ai (GLM / Zhipu)
34%±2
Win rate
Elo 882
GLM-4.6
Tier A
- Won / lost / tied
- 348 / 993 / 524
- prompts
- 14
- comparisons
- 1865
- Judge mean
- 58.4
By category
Coding48%
Creative1%
Factual37%
Reasoning25%
Rank 75OVH AI Endpoints (GRA)
32%±2
Win rate
Elo 873
- Won / lost / tied
- 572 / 1611 / 606
- prompts
- 20
- comparisons
- 2789
- Judge mean
- 74.4
By category
Coding44%
Creative9%
Factual29%
Multilingual32%
Reasoning29%
Rank 76Z.ai (GLM / Zhipu)
31%±2
Win rate
Elo 863
GLM-4.5V (vision)
Tier A
- Won / lost / tied
- 295 / 933 / 447
- prompts
- 14
- comparisons
- 1675
- Judge mean
- 54.9
By category
Coding49%
Creative8%
Factual38%
Reasoning6%
Rank 78OpenAI
29%±2
Win rate
Elo 848
gpt-5-nano-2025-08-07
Tier B
- Won / lost / tied
- 264 / 1048 / 569
- prompts
- 18
- comparisons
- 1881
- Judge mean
- 40.9
By category
Coding27%
Creative7%
Factual47%
Multilingual24%
Reasoning21%
Rank 80OVH AI Endpoints (GRA)
28%±2
Win rate
Elo 833
Qwen3.5-397B-A17B
Tier A
- Won / lost / tied
- 343 / 1299 / 485
- prompts
- 19
- comparisons
- 2127
- Judge mean
- 45.7
By category
Coding50%
Creative6%
Factual5%
Multilingual32%
Reasoning5%
Rank 80Z.ai (GLM / Zhipu)
27%±2
Win rate
Elo 831
GLM-4.6V (vision)
Tier A
- Won / lost / tied
- 249 / 1035 / 450
- prompts
- 14
- comparisons
- 1734
- Judge mean
- 47.9
By category
Coding45%
Creative4%
Factual28%
Reasoning5%
Rank 81Google Gemini
26%±2
Win rate
Elo 816
Gemini 3.5 Flash
Tier A
- Won / lost / tied
- 426 / 1808 / 555
- prompts
- 20
- comparisons
- 2789
- Judge mean
- 49.4
By category
Coding9%
Creative48%
Factual43%
Multilingual27%
Reasoning10%
Rank 82OVH AI Endpoints (GRA)
26%±2
Win rate
Elo 814
Mistral-7B-Instruct-v0.3
Tier C
- Won / lost / tied
- 463 / 1854 / 472
- prompts
- 20
- comparisons
- 2789
- Judge mean
- 70.9
By category
Coding28%
Creative12%
Factual26%
Multilingual24%
Reasoning23%
Rank 83OpenAI
24%±2
Win rate
Elo 796
gpt-5-nano
Tier C
- Won / lost / tied
- 217 / 1252 / 512
- prompts
- 20
- comparisons
- 1981
- Judge mean
- 39.4
By category
Coding18%
Creative4%
Factual50%
Multilingual18%
Reasoning10%
Rank 83Google Gemini
23%±2
Win rate
Elo 792
Nano Banana Pro
Tier A
- Won / lost / tied
- 188 / 907 / 291
- prompts
- 18
- comparisons
- 1386
- Judge mean
- 41.4
By category
Coding19%
Creative14%
Factual43%
Multilingual4%
Reasoning3%
Healthcare54%
Rank 85Google Gemini
22%±2
Win rate
Elo 780
- Won / lost / tied
- 245 / 1263 / 351
- prompts
- 24
- comparisons
- 1859
- Judge mean
- 44.1
By category
Coding4%
Creative37%
Factual42%
Multilingual20%
Reasoning13%
Healthcare3%
Rank 86Google Gemini
21%±2
Win rate
Elo 775
Gemini 3.1 Pro Preview
Tier C
- Won / lost / tied
- 417 / 2131 / 442
- prompts
- 24
- comparisons
- 2990
- Judge mean
- 47.0
By category
Coding5%
Creative50%
Factual43%
Multilingual19%
Reasoning7%
Healthcare5%
Rank 86OVH AI Endpoints (GRA)
21%±2
Win rate
Elo 772
Qwen3.5-9B
Tier B
- Won / lost / tied
- 207 / 1328 / 439
- prompts
- 19
- comparisons
- 1974
- Judge mean
- 44.8
By category
Coding37%
Creative5%
Factual10%
Multilingual24%
Reasoning4%
Rank 86Google Gemini
21%±2
Win rate
Elo 769
Gemini Pro Latest
Tier C
- Won / lost / tied
- 384 / 2127 / 479
- prompts
- 24
- comparisons
- 2990
- Judge mean
- 43.4
By category
Coding4%
Creative23%
Factual49%
Multilingual20%
Reasoning8%
Healthcare2%
Rank 88Google Gemini
19%±2
Win rate
Elo 748
Gemini 2.5 Pro
Tier A
- Won / lost / tied
- 197 / 1422 / 401
- prompts
- 24
- comparisons
- 2020
- Judge mean
- 31.3
By category
Coding17%
Creative3%
Factual34%
Multilingual15%
Reasoning3%
Healthcare54%
Rank 93Google Gemini
14%±1
Win rate
Elo 685
Gemini 2.5 Flash
Tier A
- Won / lost / tied
- 296 / 2434 / 260
- prompts
- 24
- comparisons
- 2990
- Judge mean
- 36.4
By category
Coding3%
Creative25%
Factual25%
Multilingual11%
Reasoning10%
Healthcare2%
93 models ranked from 121,133 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 160
CogView-4
Tier B
- Won / lost / tied
- 0 / 357 / 6
- prompts
- 4
- comparisons
- 363
- Judge mean
- 0.0
By category
Coding1%
Creative<1% · thin