LLMTier
Menu

Full grid / 37 products / two boards / cycle 04

The full comparison matrix

Every measured product against every axis on its own board, in two separate grids. The winning score in each axis column is marked in cyan, computed within that board's cohort only: a full-stack app builder and a coding agent are never ranked against each other in the same column, because the two boards are not comparable measurements. This is the same underlying data as the composite index and the per axis leaderboards, laid out so you can scan across a row or down a column without leaving the page. Server rendered as plain HTML: no chart library, nothing waits for JavaScript to paint.

Full-stack app builders: all 20, all 10 axes

Ordered by composite index, highest first. Cyan cell marks the column winner within this board.

Ranked leaderboard
Full-stack app builders comparison matrix: every product against every axis
#ToolAgent performanceReliabilityScalabilitySEO and GEOAPI and MCPIntegrationsDesign outputTurn latency scoreValueCode ownershipComposite
1Totalum logoTotalum
94.0●93.3●92.0●95.0●97.0●88.086.061.369.096.089.48
2Firebase Studio logoFirebase StudioSunsets Mar 2027
79.086.787.073.079.090.0●72.065.092.0●84.080.92
3Lovable logoLovable
89.082.278.075.066.086.092.076.572.079.080.54
4v0 logov0
80.084.472.076.060.066.094.0●100.0●68.086.078.31
5Replit logoReplit
83.086.780.062.071.077.073.048.072.082.075.47
6Bolt.new logoBolt.new
81.080.070.066.058.072.085.079.471.088.074.68
7Base44 logoBase44
75.084.468.064.066.075.080.069.376.055.073.05
8Anything logoAnything
71.077.862.068.057.066.083.069.376.070.069.97
9Softgen logoSoftgen
74.073.368.066.058.070.078.060.170.074.069.48
10Emergent logoEmergent
79.080.066.058.061.063.071.045.174.072.068.68
11Floot logoFloot
72.068.966.070.055.064.084.071.670.062.068.66
12Dyad logoDyad
68.055.664.072.060.062.072.077.890.098.0●67.49
13Bind AI logoBind AI
69.064.460.058.062.068.070.067.975.080.065.59
14Rocket.new logoRocket.new
70.060.061.057.052.059.075.060.866.058.062.12
15Caffeine logoCaffeine
70.066.774.048.052.046.068.043.384.058.062.08
16Zite logoZite
68.060.058.060.047.055.077.057.576.048.061.06
17bolt.diy logobolt.diy
60.055.652.062.048.054.074.057.382.097.059.85
18YouWare logoYouWare
63.048.954.061.044.052.082.081.5n/a52.058.53
19Trickle logoTrickle
62.051.152.063.041.049.079.074.974.045.058.11
20CatDoes logoCatDoes
66.048.958.042.048.060.080.047.674.066.056.99

Coding agents: all 17, all 9 axes

Ordered by composite index, highest first. Cyan cell marks the column winner within this board.

Ranked leaderboard

Code ownership has no column here: it is not scored on this board, because a coding agent edits a repository you already own.

Coding agents comparison matrix: every product against every axis
#ToolAgent performanceReliabilityScalabilitySEO and GEOAPI and MCPIntegrationsDesign outputTurn latency scoreValueComposite
1Claude Code logoClaude Code
95.0●91.186.0●62.0●78.0●76.068.047.980.079.72
2Cursor logoCursor
92.086.782.058.074.080.0●66.058.880.077.92
3Kiro logoKiro
84.093.3●80.060.072.074.065.043.478.075.67
4Augment Code logoAugment Code
86.088.984.058.074.076.062.043.977.075.63
5Amp logoAmp
85.077.874.056.072.068.060.079.177.073.35
6Factory logoFactory
81.082.282.056.076.078.062.019.476.071.48
7Antigravity logoAntigravity
83.073.376.058.070.072.070.0●39.990.0●71.37
8Zed logoZed
77.071.170.055.066.064.064.0100.0●82.071.13
9Cline logoCline
82.080.072.057.070.072.063.033.088.070.66
10Jules logoJules
80.075.674.055.066.070.062.021.488.068.04
11Devin logoDevin
78.071.180.054.072.074.060.08.974.066.64
12Qodo logoQodo
70.077.876.052.068.070.056.031.670.066.06
13Aider logoAider
72.068.964.054.062.058.052.079.786.065.88
14Codebuff logoCodebuff
76.068.968.055.064.062.058.052.552.064.51
15OpenHands logoOpenHands
74.064.470.056.068.066.058.021.786.063.93
16Trae logoTrae
73.051.166.054.060.066.066.032.584.061.00
17Roomote logoRoomote
71.053.366.053.062.060.056.032.684.059.89

Column winners

The single highest subscore recorded on each axis, within each board.

Full-stack app builders

Agent performance

Totalum logoTotalum

94.0

Reliability

Totalum logoTotalum

93.3

Scalability

Totalum logoTotalum

92.0

SEO and GEO

Totalum logoTotalum

95.0

API and MCP

Totalum logoTotalum

97.0

Integrations

Firebase Studio logoFirebase StudioSunsets Mar 2027

90.0

Design output

v0 logov0

94.0

Turn latency score

v0 logov0

100.0

Value

Firebase Studio logoFirebase StudioSunsets Mar 2027

92.0

Code ownership

Dyad logoDyad

98.0

Coding agents

Agent performance

Claude Code logoClaude Code

95.0

Reliability

Kiro logoKiro

93.3

Scalability

Claude Code logoClaude Code

86.0

SEO and GEO

Claude Code logoClaude Code

62.0

API and MCP

Claude Code logoClaude Code

78.0

Integrations

Cursor logoCursor

80.0

Design output

Antigravity logoAntigravity

70.0

Turn latency score

Zed logoZed

100.0

Value

Antigravity logoAntigravity

90.0

Reading the grid

Every number here is the same subscore published on the product page and used in the weighted composite on the methodology page. Nothing is normalised again for this view: a 98 in one column and a 98 in another are the same measurement scale, 0 to 100, within the same board. A 98 on the builders board and a 98 on the coding agent board are not necessarily the same thing, because the two boards score different axis sets with different weights.

The composite column on the right is the weighted sum of that board's axis columns using its own published weights, so a product that wins several low weight axes will not necessarily outrank one that wins fewer high weight axes. Agent performance and reliability together carry 36 of the 100 points on the builders board, and 37.11 of the 100 points on the coding agent board, more than the next four axes combined on either board.

The full underlying dataset, including every raw run behind these subscores, is downloadable at /api/rankings.json and /api/runs.json.