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.

How this is scored

8,490

scored runs

80,516

pairwise comparisons

24

prompts

4,422

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 1161

gpt-5.1
Tier B
Won / lost / tied
985 / 163 / 1188
prompts
20
comparisons
2336
Judge mean
92.9

By category

Coding79%
Creative77%
Factual66%
Multilingual65%
Reasoning80%
Rank 1OpenAI
71%±2

Win rate

Elo 1158

gpt-5.5-2026-04-23
Tier A
Won / lost / tied
1010 / 203 / 1123
prompts
20
comparisons
2336
Judge mean
93.9

By category

Coding80%
Creative89%
Factual75%
Multilingual62%
Reasoning65%
Rank 1OpenAI
71%±2

Win rate

Elo 1158

gpt-4.1-mini
Tier C
Won / lost / tied
1072 / 216 / 1194
prompts
24
comparisons
2482
Judge mean
94.6

By category

Coding76%
Creative89%
Factual70%
Multilingual62%
Reasoning78%
Healthcare67%
Rank 1OpenAI
71%±3

Win rate

Elo 1155

gpt-5.6-terra
Tier B
Won / lost / tied
316 / 92 / 291
prompts
10
comparisons
699
Judge mean
89.5

By category

Coding76%
Factual61%
Reasoning78%
Rank 1OpenAI
71%±2

Win rate

Elo 1154

gpt-5.5
Tier C
Won / lost / tied
988 / 204 / 1144
prompts
20
comparisons
2336
Judge mean
92.6

By category

Coding74%
Creative85%
Factual76%
Multilingual59%
Reasoning76%
Rank 1OpenAI
70%±2

Win rate

Elo 1149

gpt-4.1-2025-04-14
Tier C
Won / lost / tied
1000 / 196 / 1286
prompts
24
comparisons
2482
Judge mean
94.8

By category

Coding77%
Creative91%
Factual69%
Multilingual62%
Reasoning70%
Healthcare76%
Rank 1OpenAI
70%±2

Win rate

Elo 1144

gpt-5.1-2025-11-13
Tier B
Won / lost / tied
928 / 203 / 1205
prompts
20
comparisons
2336
Judge mean
92.5

By category

Coding79%
Creative74%
Factual64%
Multilingual63%
Reasoning76%
Rank 1OpenAI
69%±2

Win rate

Elo 1142

gpt-5.2
Tier B
Won / lost / tied
942 / 227 / 1167
prompts
20
comparisons
2336
Judge mean
91.6

By category

Coding64%
Creative87%
Factual68%
Multilingual66%
Reasoning79%
Rank 1OpenAI
69%±2

Win rate

Elo 1137

gpt-5.4-2026-03-05
Tier B
Won / lost / tied
927 / 241 / 1168
prompts
20
comparisons
2336
Judge mean
93.0

By category

Coding76%
Creative89%
Factual52%
Multilingual66%
Reasoning76%
Rank 1OpenAI
68%±2

Win rate

Elo 1134

gpt-5.2-2025-12-11
Tier B
Won / lost / tied
902 / 235 / 1199
prompts
20
comparisons
2336
Judge mean
91.7

By category

Coding73%
Creative79%
Factual53%
Multilingual68%
Reasoning80%
Rank 1Google Gemini
68%±4

Win rate

Elo 1133

Gemini 3.5 Flash-Lite
Tier B
Won / lost / tied
284 / 117 / 227
prompts
9
comparisons
628
Judge mean
84.1

By category

Coding70%
Factual74%
Reasoning65%
Rank 1Anthropic
68%±2

Win rate

Elo 1133

Claude Opus 4.6
Tier B
Won / lost / tied
980 / 300 / 1102
prompts
24
comparisons
2382
Judge mean
96.3

By category

Coding65%
Creative92%
Factual72%
Multilingual63%
Reasoning67%
Healthcare85%
Rank 4OpenAI
67%±2

Win rate

Elo 1125

gpt-4.1-mini-2025-04-14
Tier C
Won / lost / tied
936 / 279 / 1267
prompts
24
comparisons
2482
Judge mean
93.5

By category

Coding72%
Creative84%
Factual60%
Multilingual61%
Reasoning75%
Healthcare62%
Rank 6OpenAI
66%±2

Win rate

Elo 1118

gpt-5.4
Tier A
Won / lost / tied
875 / 302 / 1159
prompts
20
comparisons
2336
Judge mean
91.7

By category

Coding73%
Creative83%
Factual50%
Multilingual63%
Reasoning76%
Rank 3OpenAI
66%±4

Win rate

Elo 1115

gpt-5.6-luna
Tier B
Won / lost / tied
285 / 131 / 283
prompts
10
comparisons
699
Judge mean
87.4

By category

Coding64%
Factual61%
Reasoning78%
Rank 3OpenAI
66%±4

Win rate

Elo 1115

gpt-6-astra
Tier A
Won / lost / tied
326 / 172 / 201
prompts
10
comparisons
699
Judge mean
89.1

By category

Coding78%
Factual76%
Reasoning39%
Rank 6OpenAI
66%±2

Win rate

Elo 1114

gpt-4.1
Tier B
Won / lost / tied
920 / 329 / 1233
prompts
24
comparisons
2482
Judge mean
94.0

By category

Coding72%
Creative81%
Factual67%
Multilingual61%
Reasoning63%
Healthcare70%
Rank 9OpenAI
66%±2

Win rate

Elo 1112

gpt-5.4-mini-2026-03-17
Tier A
Won / lost / tied
859 / 323 / 1154
prompts
20
comparisons
2336
Judge mean
89.7

By category

Coding65%
Creative89%
Factual53%
Multilingual61%
Reasoning79%
Rank 5OpenAI
65%±4

Win rate

Elo 1109

gpt-6-luna
Tier B
Won / lost / tied
325 / 182 / 192
prompts
10
comparisons
699
Judge mean
88.7

By category

Coding66%
Factual84%
Reasoning46%
Rank 5OpenAI
65%±4

Win rate

Elo 1109

gpt-6.1-sol
Tier A
Won / lost / tied
314 / 171 / 214
prompts
10
comparisons
699
Judge mean
88.2

By category

Coding70%
Factual67%
Reasoning61%
Rank 9OpenAI
65%±2

Win rate

Elo 1109

gpt-5.4-mini
Tier A
Won / lost / tied
875 / 356 / 1105
prompts
20
comparisons
2336
Judge mean
90.7

By category

Coding71%
Creative86%
Factual55%
Multilingual64%
Reasoning66%
Rank 9Anthropic
65%±2

Win rate

Elo 1107

Claude Opus 4.7
Tier B
Won / lost / tied
899 / 371 / 1112
prompts
24
comparisons
2382
Judge mean
94.4

By category

Coding51%
Creative60%
Factual79%
Multilingual63%
Reasoning75%
Healthcare73%
Rank 12OpenAI
64%±2

Win rate

Elo 1099

o3
Tier C
Won / lost / tied
825 / 369 / 1092
prompts
20
comparisons
2286
Judge mean
85.6

By category

Coding69%
Creative86%
Factual55%
Multilingual61%
Reasoning65%
Rank 12OpenAI
64%±2

Win rate

Elo 1098

gpt-4o
Tier C
Won / lost / tied
900 / 410 / 1172
prompts
24
comparisons
2482
Judge mean
93.3

By category

Coding58%
Creative68%
Factual60%
Multilingual65%
Reasoning75%
Healthcare60%
Rank 13Anthropic
63%±2

Win rate

Elo 1093

Claude Opus 4.5
Tier B
Won / lost / tied
874 / 437 / 1071
prompts
24
comparisons
2382
Judge mean
95.0

By category

Coding58%
Creative60%
Factual69%
Multilingual62%
Reasoning67%
Healthcare83%
Rank 13Google Gemini
63%±2

Win rate

Elo 1092

Gemini Flash-Lite Latest
Tier C
Won / lost / tied
911 / 464 / 1107
prompts
24
comparisons
2482
Judge mean
93.5

By category

Coding70%
Creative51%
Factual67%
Multilingual59%
Reasoning63%
Healthcare78%
Rank 15OpenAI
62%±2

Win rate

Elo 1088

gpt-5-search-api
Tier C
Won / lost / tied
868 / 447 / 1167
prompts
24
comparisons
2482
Judge mean
92.5

By category

Coding64%
Creative83%
Factual52%
Multilingual56%
Reasoning74%
Healthcare82%
Rank 15Anthropic
62%±2

Win rate

Elo 1087

Claude Fable 5
Tier A
Won / lost / tied
795 / 416 / 1055
prompts
20
comparisons
2266
Judge mean
87.5

By category

Coding66%
Creative68%
Factual57%
Multilingual59%
Reasoning70%
Rank 15OpenAI
62%±2

Win rate

Elo 1086

gpt-4o-2024-08-06
Tier C
Won / lost / tied
873 / 464 / 1145
prompts
24
comparisons
2482
Judge mean
91.2

By category

Coding53%
Creative72%
Factual69%
Multilingual63%
Reasoning66%
Healthcare66%
Rank 16OpenAI
62%±2

Win rate

Elo 1083

o3-2025-04-16
Tier B
Won / lost / tied
774 / 427 / 1016
prompts
20
comparisons
2217
Judge mean
85.9

By category

Coding66%
Creative86%
Factual52%
Multilingual58%
Reasoning65%
Rank 17OpenAI
62%±2

Win rate

Elo 1082

gpt-5-search-api-2025-10-14
Tier B
Won / lost / tied
864 / 481 / 1137
prompts
24
comparisons
2482
Judge mean
92.0

By category

Coding59%
Creative77%
Factual56%
Multilingual52%
Reasoning78%
Healthcare83%
Rank 18Anthropic
61%±2

Win rate

Elo 1078

Claude Sonnet 4.5
Tier B
Won / lost / tied
823 / 476 / 1083
prompts
24
comparisons
2382
Judge mean
93.4

By category

Coding43%
Creative59%
Factual68%
Multilingual61%
Reasoning78%
Healthcare69%
Rank 14OpenAI
61%±4

Win rate

Elo 1077

gpt-5.6-sol
Tier B
Won / lost / tied
294 / 210 / 195
prompts
10
comparisons
699
Judge mean
86.4

By category

Coding65%
Factual60%
Reasoning61%
Rank 19Google Gemini
61%±2

Win rate

Elo 1076

Gemini 3.1 Flash Lite
Tier B
Won / lost / tied
757 / 435 / 1144
prompts
20
comparisons
2336
Judge mean
89.6

By category

Coding55%
Creative40%
Factual69%
Multilingual60%
Reasoning70%
Rank 19Z.ai (GLM / Zhipu)
61%±2

Win rate

Elo 1075

GLM-4.7
Tier A
Won / lost / tied
630 / 389 / 649
prompts
14
comparisons
1668
Judge mean
89.1

By category

Coding55%
Creative83%
Factual63%
Reasoning63%
Rank 19OpenAI
60%±2

Win rate

Elo 1074

gpt-4o-2024-11-20
Tier C
Won / lost / tied
818 / 486 / 1178
prompts
24
comparisons
2482
Judge mean
92.8

By category

Coding60%
Creative51%
Factual62%
Multilingual58%
Reasoning68%
Healthcare78%
Rank 19Google Gemini
60%±3

Win rate

Elo 1072

Nano Banana
Tier Barchived
Won / lost / tied
442 / 232 / 692
prompts
23
comparisons
1366
Judge mean
93.5

By category

Coding51%
Creative66%
Factual56%
Multilingual63%
Reasoning74%
Healthcare50%
Rank 21Anthropic
60%±2

Win rate

Elo 1071

Claude Opus 4.8
Tier A
Won / lost / tied
767 / 487 / 982
prompts
20
comparisons
2236
Judge mean
92.7

By category

Coding54%
Creative39%
Factual68%
Multilingual63%
Reasoning67%
Rank 21OVH AI Endpoints (GRA)
60%±2

Win rate

Elo 1070

gpt-oss-120b
Tier C
Won / lost / tied
788 / 504 / 1044
prompts
20
comparisons
2336
Judge mean
89.8

By category

Coding63%
Creative83%
Factual57%
Multilingual55%
Reasoning60%
Rank 25Anthropic
59%±2

Win rate

Elo 1066

Claude Sonnet 4.6
Tier A
Won / lost / tied
778 / 503 / 1101
prompts
24
comparisons
2382
Judge mean
93.6

By category

Coding61%
Creative55%
Factual63%
Multilingual54%
Reasoning67%
Healthcare75%
Rank 23Anthropic
59%±3

Win rate

Elo 1064

Claude Sonnet 5
Tier A
Won / lost / tied
556 / 415 / 525
prompts
19
comparisons
1496
Judge mean
84.0

By category

Coding60%
Creative84%
Factual48%
Multilingual58%
Reasoning66%
Rank 29OpenAI
58%±2

Win rate

Elo 1057

gpt-4o-mini
Tier C
Won / lost / tied
779 / 558 / 1145
prompts
24
comparisons
2482
Judge mean
87.4

By category

Coding61%
Creative65%
Factual38%
Multilingual62%
Reasoning72%
Healthcare49%
Rank 30OpenAI
58%±2

Win rate

Elo 1054

gpt-5.4-nano-2026-03-17
Tier A
Won / lost / tied
765 / 581 / 990
prompts
20
comparisons
2336
Judge mean
83.9

By category

Coding68%
Creative30%
Factual39%
Multilingual65%
Reasoning66%
Rank 27Google Gemini
57%±3

Win rate

Elo 1052

Nano Banana 2 (preview)
Tier B
Won / lost / tied
362 / 239 / 470
prompts
18
comparisons
1071
Judge mean
92.6

By category

Coding71%
Creative60%
Factual47%
Multilingual56%
Reasoning64%
Healthcare35%
Rank 32OpenAI
57%±2

Win rate

Elo 1049

gpt-5-mini
Tier C
Won / lost / tied
634 / 487 / 989
prompts
20
comparisons
2110
Judge mean
83.6

By category

Coding68%
Creative68%
Factual47%
Multilingual53%
Reasoning58%
Rank 33Z.ai (GLM / Zhipu)
56%±2

Win rate

Elo 1044

GLM-4.5
Tier A
Won / lost / tied
598 / 491 / 579
prompts
14
comparisons
1668
Judge mean
86.1

By category

Coding48%
Creative84%
Factual68%
Reasoning45%
Rank 42OpenAI
54%±2

Win rate

Elo 1030

gpt-4o-mini-2024-07-18
Tier C
Won / lost / tied
734 / 692 / 1056
prompts
24
comparisons
2482
Judge mean
86.6

By category

Coding61%
Creative65%
Factual40%
Multilingual51%
Reasoning66%
Healthcare46%
Rank 43OpenAI
54%±2

Win rate

Elo 1025

gpt-5.4-nano
Tier C
Won / lost / tied
768 / 758 / 810
prompts
20
comparisons
2336
Judge mean
82.2

By category

Coding58%
Creative32%
Factual24%
Multilingual70%
Reasoning66%
Rank 43OpenAI
53%±2

Win rate

Elo 1024

gpt-5-mini-2025-08-07
Tier B
Won / lost / tied
619 / 613 / 919
prompts
20
comparisons
2151
Judge mean
81.3

By category

Coding69%
Creative36%
Factual44%
Multilingual49%
Reasoning59%
Rank 44OVH AI Endpoints (GRA)
53%±2

Win rate

Elo 1023

Qwen2.5-VL-72B-Instruct
Tier B
Won / lost / tied
714 / 716 / 906
prompts
20
comparisons
2336
Judge mean
88.4

By category

Coding56%
Creative64%
Factual28%
Multilingual54%
Reasoning76%
Rank 41OpenAI
53%±4

Win rate

Elo 1021

gpt-6-sol
Tier A
Won / lost / tied
244 / 263 / 192
prompts
10
comparisons
699
Judge mean
86.0

By category

Coding70%
Factual45%
Reasoning39%
Rank 43Z.ai (GLM / Zhipu)
53%±3

Win rate

Elo 1021

GLM-5
Tier A
Won / lost / tied
432 / 409 / 582
prompts
14
comparisons
1423
Judge mean
88.2

By category

Coding52%
Creative87%
Factual56%
Reasoning44%
Rank 44Anthropic
53%±2

Win rate

Elo 1020

Claude Haiku 4.5
Tier A
Won / lost / tied
705 / 721 / 956
prompts
24
comparisons
2382
Judge mean
90.8

By category

Coding63%
Creative41%
Factual49%
Multilingual48%
Reasoning61%
Healthcare30%
Rank 44OVH AI Endpoints (GRA)
53%±2

Win rate

Elo 1019

Meta-Llama-3_3-70B-Instruct
Tier B
Won / lost / tied
651 / 679 / 1006
prompts
20
comparisons
2336
Judge mean
88.1

By category

Coding45%
Creative73%
Factual49%
Multilingual41%
Reasoning76%
Rank 46Anthropic
52%±2

Win rate

Elo 1013

Claude Opus 5
Tier A
Won / lost / tied
558 / 614 / 672
prompts
20
comparisons
1844
Judge mean
83.2

By category

Coding33%
Creative87%
Factual48%
Multilingual52%
Reasoning55%
Rank 47OVH AI Endpoints (GRA)
52%±2

Win rate

Elo 1012

Mistral-Small-3.2-24B-Instruct-2506
Tier Barchived
Won / lost / tied
690 / 765 / 881
prompts
20
comparisons
2336
Judge mean
87.9

By category

Coding69%
Creative41%
Factual40%
Multilingual46%
Reasoning55%
Rank 47Google Gemini
51%±3

Win rate

Elo 1006

Gemini Robotics-ER 1.6 Preview
Tier Barchived
Won / lost / tied
373 / 398 / 595
prompts
23
comparisons
1366
Judge mean
79.3

By category

Coding47%
Creative80%
Factual40%
Multilingual56%
Reasoning46%
Healthcare41%
Rank 47Google Gemini
50%±2

Win rate

Elo 1001

Gemini 2.5 Flash-Lite
Tier B
Won / lost / tied
750 / 893 / 839
prompts
24
comparisons
2482
Judge mean
88.9

By category

Coding40%
Creative40%
Factual40%
Multilingual59%
Reasoning64%
Healthcare48%
Rank 54Google Gemini
46%±4

Win rate

Elo 973

Gemini 3.7 Flash
Tier B
Won / lost / tied
193 / 290 / 145
prompts
9
comparisons
628
Judge mean
66.0

By category

Coding44%
Factual60%
Reasoning34%
Rank 59Google Gemini
45%±2

Win rate

Elo 963

Gemini 3 Flash Preview
Tier C
Won / lost / tied
686 / 1074 / 722
prompts
24
comparisons
2482
Judge mean
77.9

By category

Coding33%
Creative22%
Factual51%
Multilingual50%
Reasoning52%
Healthcare25%
Rank 59OpenAI
44%±2

Win rate

Elo 960

gpt-5-2025-08-07
Tier B
Won / lost / tied
482 / 791 / 709
prompts
20
comparisons
1982
Judge mean
70.8

By category

Coding57%
Creative3%
Factual20%
Multilingual50%
Reasoning58%
Rank 59OpenAI
44%±2

Win rate

Elo 957

gpt-5
Tier C
Won / lost / tied
430 / 731 / 682
prompts
19
comparisons
1843
Judge mean
73.5

By category

Coding58%
Creative7%
Factual42%
Multilingual39%
Reasoning47%
Rank 63OVH AI Endpoints (GRA)
36%±2

Win rate

Elo 898

gpt-oss-20b
Tier C
Won / lost / tied
470 / 1169 / 581
prompts
20
comparisons
2220
Judge mean
80.8

By category

Coding50%
Creative42%
Factual17%
Multilingual26%
Reasoning49%
Rank 63Z.ai (GLM / Zhipu)
36%±3

Win rate

Elo 897

GLM-4.5 Air
Tier B
Won / lost / tied
324 / 793 / 435
prompts
14
comparisons
1552
Judge mean
63.9

By category

Coding54%
Creative44%
Factual22%
Reasoning20%
Rank 63Google Gemini
35%±4

Win rate

Elo 892

Gemini 3.8 Flash
Tier A
Won / lost / tied
143 / 364 / 121
prompts
9
comparisons
628
Judge mean
56.5

By category

Coding25%
Factual47%
Reasoning32%
Rank 63Z.ai (GLM / Zhipu)
34%±3

Win rate

Elo 882

GLM-5 Turbo
Tier B
Won / lost / tied
257 / 750 / 453
prompts
14
comparisons
1460
Judge mean
60.4

By category

Coding44%
Creative32%
Factual39%
Reasoning11%
Rank 63Google Gemini
34%±2

Win rate

Elo 881

Gemini Flash Latest
Tier B
Won / lost / tied
538 / 1418 / 526
prompts
24
comparisons
2482
Judge mean
63.6

By category

Coding32%
Creative40%
Factual38%
Multilingual30%
Reasoning32%
Healthcare14%
Rank 63OpenAI
32%±2

Win rate

Elo 873

gpt-5-nano-2025-08-07
Tier B
Won / lost / tied
252 / 807 / 483
prompts
19
comparisons
1542
Judge mean
53.6

By category

Coding27%
Creative7%
Factual49%
Multilingual28%
Reasoning27%
Rank 63Z.ai (GLM / Zhipu)
32%±3

Win rate

Elo 867

GLM-5.1
Tier A
Won / lost / tied
285 / 807 / 331
prompts
14
comparisons
1423
Judge mean
68.8

By category

Coding28%
Creative15%
Factual49%
Reasoning22%
Rank 65Z.ai (GLM / Zhipu)
30%±2

Win rate

Elo 855

GLM-4.5V (vision)
Tier A
Won / lost / tied
270 / 861 / 362
prompts
14
comparisons
1493
Judge mean
63.3

By category

Coding42%
Creative9%
Factual40%
Reasoning10%
Rank 65Z.ai (GLM / Zhipu)
30%±2

Win rate

Elo 851

GLM-4.6
Tier A
Won / lost / tied
257 / 894 / 394
prompts
14
comparisons
1545
Judge mean
65.1

By category

Coding38%
Creative1%
Factual31%
Reasoning28%
Rank 67OVH AI Endpoints (GRA)
29%±2

Win rate

Elo 843

Qwen3.5-397B-A17B
Tier A
Won / lost / tied
302 / 1043 / 389
prompts
19
comparisons
1734
Judge mean
50.7

By category

Coding53%
Creative6%
Factual4%
Multilingual36%
Reasoning5%
Rank 66Z.ai (GLM / Zhipu)
29%±2

Win rate

Elo 842

GLM-4.6V (vision)
Tier A
Won / lost / tied
209 / 812 / 402
prompts
14
comparisons
1423
Judge mean
54.6

By category

Coding54%
Creative4%
Factual16%
Reasoning19%
Rank 70Google Gemini
26%±2

Win rate

Elo 822

Gemini 3.5 Flash
Tier A
Won / lost / tied
376 / 1504 / 456
prompts
20
comparisons
2336
Judge mean
49.6

By category

Coding9%
Creative53%
Factual45%
Multilingual28%
Reasoning11%
Rank 70OpenAI
26%±2

Win rate

Elo 821

gpt-5-nano
Tier C
Won / lost / tied
176 / 905 / 477
prompts
20
comparisons
1558
Judge mean
52.1

By category

Coding18%
Creative4%
Factual43%
Multilingual25%
Reasoning18%
Rank 71Z.ai (GLM / Zhipu)
26%±2

Win rate

Elo 814

GLM-5.2
Tier A
Won / lost / tied
238 / 1055 / 381
prompts
14
comparisons
1674
Judge mean
52.5

By category

Coding35%
Creative4%
Factual37%
Reasoning6%
Rank 74Google Gemini
23%±2

Win rate

Elo 790

Gemini 3.1 Pro Preview Custom Tools
Tier C
Won / lost / tied
205 / 948 / 271
prompts
24
comparisons
1424
Judge mean
44.1

By category

Coding5%
Creative35%
Factual38%
Multilingual22%
Reasoning17%
Healthcare6%
Rank 74Google Gemini
23%±3

Win rate

Elo 786

Nano Banana Pro (preview)
Tier A
Won / lost / tied
136 / 692 / 243
prompts
18
comparisons
1071
Judge mean
41.4

By category

Coding19%
Creative13%
Factual39%
Multilingual5%
Reasoning4%
Healthcare53%
Rank 75OVH AI Endpoints (GRA)
22%±2

Win rate

Elo 780

Qwen3.5-9B
Tier B
Won / lost / tied
184 / 1099 / 376
prompts
19
comparisons
1659
Judge mean
48.2

By category

Coding37%
Creative5%
Factual9%
Multilingual25%
Reasoning4%
Rank 77Google Gemini
22%±2

Win rate

Elo 776

Gemini 3.1 Pro Preview
Tier C
Won / lost / tied
364 / 1777 / 341
prompts
24
comparisons
2482
Judge mean
47.5

By category

Coding7%
Creative53%
Factual41%
Multilingual18%
Reasoning9%
Healthcare9%
Rank 77Google Gemini
19%±2

Win rate

Elo 753

Gemini Pro Latest
Tier C
Won / lost / tied
317 / 1830 / 335
prompts
24
comparisons
2482
Judge mean
43.2

By category

Coding5%
Creative27%
Factual40%
Multilingual19%
Reasoning8%
Healthcare5%
Rank 78Google Gemini
19%±2

Win rate

Elo 744

Gemini 2.5 Pro
Tier A
Won / lost / tied
132 / 1078 / 347
prompts
24
comparisons
1557
Judge mean
34.1

By category

Coding16%
Creative3%
Factual29%
Multilingual16%
Reasoning4%
Healthcare53%
Rank 83Google Gemini
15%±1

Win rate

Elo 698

Gemini 2.5 Flash
Tier A
Won / lost / tied
260 / 1984 / 238
prompts
24
comparisons
2482
Judge mean
35.5

By category

Coding4%
Creative23%
Factual21%
Multilingual16%
Reasoning11%
Healthcare4%
Rank 82Google Gemini
14%±3

Win rate

Elo 691

Gemini 3.6 Flash
Tier B
Won / lost / tied
55 / 507 / 66
prompts
9
comparisons
628
Judge mean
43.9

By category

Coding2%
Factual31%
Reasoning9%

84 models ranked from 80,516 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 195

CogView-4
Tier B
Won / lost / tied
0 / 277 / 6
prompts
4
comparisons
283
Judge mean
0.0

By category

Coding1%
Creative<1% · thin