Model benchmark rankings

Compare capability scores, evaluation time, and reference cost across models and reasoning efforts.

Evaluated on private datasetsWhy we test this way

Data snapshot · 10/20/2026

Scoring dimension
About the scores

Backend engineering 20% · Frontend interaction 20% · Reasoning 35% · Aesthetics 15% · Maintainability 10%. By default, each model shows its highest-scoring tested effort for the selected dimension.

Evaluation time and reference cost correspond to the selected dimension. They reflect benchmark runs, not everyday single-request latency or billing.

Explore evaluation tasksEvaluation method
Filters
Result scope
Rank byBy score
Ranking method

Preference rankings select the tested effort suited to that method; raw scores stay unchanged.
These are benchmark completion times, not everyday chat response speeds.

42 models · 87 configsBest effort per model

Scroll horizontally to view all scores and costs →

RankModelEval. time
Speed tok/sActions
01
claude-opus-5-5
Backend engineering84Frontend interaction97Knowledge & Reasoning100Visual aesthetics73Maintainability70
89.158497100737066m 3s$13.41…
02
gpt-6-astra
Backend engineering95Frontend interaction96Knowledge & Reasoning95Visual aesthetics75Maintainability61
88.80959695756183m 56s$7.01…
03
claude-fable-5-1
Backend engineering85Frontend interaction91Knowledge & Reasoning95Visual aesthetics80Maintainability68
87.25859195806879m 7s$26.15…
04
gpt-6.1-sol
Backend engineering95Frontend interaction96Knowledge & Reasoning90Visual aesthetics66Maintainability70
86.60959690667062m 9s$1.69…
05
claude-sonnet-5-5
Backend engineering86Frontend interaction89Knowledge & Reasoning85Visual aesthetics73Maintainability66
82.30868985736648m 32s$2.80…
06
claude-opus-5
Backend engineering70Frontend interaction87Knowledge & Reasoning95Visual aesthetics73Maintainability61
81.70708795736192m 12s$7.99…
07
muse-wonderNEW
Backend engineering82Frontend interaction84Knowledge & Reasoning90Visual aesthetics73Maintainability59
81.55828490735940m 8s$6.78…
08
Gemini 4 ArgonNEW
Backend engineering82Frontend interaction84Knowledge & Reasoning85Visual aesthetics70Maintainability59
79.35828485705946m 48sN/A…
09
gpt-6-sol
Backend engineering80Frontend interaction94Knowledge & Reasoning75Visual aesthetics75Maintainability66
78.90809475756639m 22s$1.10…
10
grok-4.7
Backend engineering87Frontend interaction80Knowledge & Reasoning85Visual aesthetics66Maintainability55
78.558780856655196m 18s$5.25…
11
grok-4.6
Backend engineering78Frontend interaction80Knowledge & Reasoning85Visual aesthetics64Maintainability59
76.857880856459144m 16s$2.51…
12
claude-opus-4-8
Backend engineering61Frontend interaction87Knowledge & Reasoning90Visual aesthetics68Maintainability55
76.806187906855102m 10s$9.18…
13
deepseek-v4.1-flash-1017NEW
Backend engineering72Frontend interaction96Knowledge & Reasoning75Visual aesthetics61Maintainability53
74.30729675615364m 50s$0.45…
14
gpt-5.6-sol
Backend engineering85Frontend interaction73Knowledge & Reasoning75Visual aesthetics66Maintainability64
74.15857375666487m 21s$3.25…
15
MiniMax-M3.1-FlashNEW
Backend engineering73Frontend interaction75Knowledge & Reasoning80Visual aesthetics66Maintainability61
73.60737580666187m 15s$0.21…
16
muse-spark-1.3
Backend engineering73Frontend interaction91Knowledge & Reasoning75Visual aesthetics59Maintainability47
72.60739175594729m 10s$0.066…
17
k3
Backend engineering70Frontend interaction85Knowledge & Reasoning70Visual aesthetics68Maintainability59
71.607085706859158m 17s$3.28…
18
glm-5.3
Backend engineering70Frontend interaction78Knowledge & Reasoning70Visual aesthetics66Maintainability64
70.407078706664153m 15s$1.83…
19
mimo-v2.6-pro
Backend engineering70Frontend interaction85Knowledge & Reasoning65Visual aesthetics64Maintainability59
69.257085656459336m 3s$0.66…
20
space-bunny
Backend engineering61Frontend interaction80Knowledge & Reasoning75Visual aesthetics59Maintainability55
68.80618075595562m 35s$0.000…
21
MiniMax-M3.1-Flash-Preview
Backend engineering75Frontend interaction57Knowledge & Reasoning75Visual aesthetics70Maintainability53
68.45755775705354m 7sN/A…
22
deepseek-v4.1-flash
Backend engineering62Frontend interaction88Knowledge & Reasoning65Visual aesthetics64Maintainability55
67.85628865645535m 40s$0.41…
23
hy4-preview
Backend engineering74Frontend interaction73Knowledge & Reasoning65Visual aesthetics59Maintainability57
66.707473655957218m 41s$1.63…
24
gpt-5.6-terra
Backend engineering81Frontend interaction76Knowledge & Reasoning55Visual aesthetics61Maintainability61
65.908176556161133m 2s$1.94…
25
claude-haiku-5-5
Backend engineering69Frontend interaction78Knowledge & Reasoning60Visual aesthetics59Maintainability64
65.65697860596427m 38s$0.21…
26
step-5
Backend engineering61Frontend interaction71Knowledge & Reasoning75Visual aesthetics47Maintainability55
65.20617175475598m 22s$1.24…
27
deepseek-v4-pro
Backend engineering67Frontend interaction63Knowledge & Reasoning65Visual aesthetics70Maintainability59
65.15676365705993m 15s$1.15…
27
Qwen 3.8 Flash
Backend engineering68Frontend interaction87Knowledge & Reasoning50Visual aesthetics73Maintainability57
65.156887507357134m 31s$0.42…
29
k2.8
Backend engineering54Frontend interaction72Knowledge & Reasoning70Visual aesthetics57Maintainability53
63.55547270575370m 58s$0.66…
30
gpt-6-luna
Backend engineering62Frontend interaction80Knowledge & Reasoning50Visual aesthetics68Maintainability59
62.00628050685963m 52s$0.061…
31
doubao-seed-2-1-pro
Backend engineering68Frontend interaction68Knowledge & Reasoning55Visual aesthetics68Maintainability49
61.556868556849104m 54s$1.38…
32
deepseek-v4-flash
Backend engineering58Frontend interaction76Knowledge & Reasoning55Visual aesthetics61Maintainability53
60.50587655615398m 25s$0.42…
33
deepseek-v4-flash-vision-exp
Backend engineering61Frontend interaction43Knowledge & Reasoning70Visual aesthetics61Maintainability55
59.95614370615588m 43s$0.32…
34
Gemini 3.8 Flash High
Backend engineering64Frontend interaction86Knowledge & Reasoning40Visual aesthetics66Maintainability59
59.80648640665925m 21s$0.13…
35
glm-5.3-flash
Backend engineering54Frontend interaction75Knowledge & Reasoning40Visual aesthetics78Maintainability53
56.805475407853102m 55s$0.081…
36
mimo-v2.6-flash
Backend engineering58Frontend interaction71Knowledge & Reasoning45Visual aesthetics59Maintainability57
56.105871455957386m 14s$0.28…
37
Qwen 3.8 Max
Backend engineering64Frontend interaction69Knowledge & Reasoning40Visual aesthetics64Maintainability53
55.506469406453133m 3s$3.41…
38
Qwen 3.8 27b
Backend engineering64Frontend interaction87Knowledge & Reasoning20Visual aesthetics66Maintainability59
53.006487206659104m 31s$1.79…
39
MiniMax-M3
Backend engineering46Frontend interaction65Knowledge & Reasoning50Visual aesthetics59Maintainability44
52.95466550594456m 33s$0.62…
40
gpt-5.6-luna
Backend engineering56Frontend interaction83Knowledge & Reasoning25Visual aesthetics66Maintainability59
52.35568325665975m 23s$0.16…
41
Gemini 3.7 Flash High
Backend engineering63Frontend interaction71Knowledge & Reasoning25Visual aesthetics70Maintainability59
51.95637125705924m 19s$0.13…
42
Nemotron-3-Ultra
Backend engineering46Frontend interaction40Knowledge & Reasoning40Visual aesthetics61Maintainability37
44.054640406137109m 30s$0.000…
Each model uses its highest-scoring effort here. Tied scores share a rank.
Backend engineering 20% · Frontend interaction 20% · Reasoning 35% · Aesthetics 15% · Maintainability 10%. Evaluation scope and sources