How the builds compare

#01 Claude Code Team

Claude Code v2.1.39Team ModeOpus 4.6

View full report →

#02 Claude Code Sub-Agents

Claude Code v2.1.41Sub-AgentsOpus 4.6

View full report →

#03 Codex Sub-Agents

Codex v0.99.0Sub-AgentsGPT-5.3-codex

View full report →

#04 Claude Code Team v2

Claude Code v2.1.41Sub-AgentsOpus 4.6

View full report →

#05 Codex Sub-Agents v2

Codex v0.101.0Sub-AgentsGPT-5.3-codexQuality

View full report →

#06 Claude Code Team v3

Claude Code v2.1.80Team ModeOpus 4.61M Context

View full report →

#07 Claude Code Team v4

Claude Code v2.1.81Team ModeOpus 4.61M Context

View full report →

Feature Completeness (143 E2E Tests, verified by 2 independent QA Analysis)

Fully WorkingAll acceptance criteria met

126 (88%)

73 (51%)

89 (62%)

82 (57%)

70 (49%)

119 (83%)

121 (85%)

Not WorkingFeature missing or broken

9

43

39

30

39

7

7

Partially DoneSome criteria met, gaps remain

7

27

10

27

26

13

15

Not ApplicableFeature outside build scope

1

0

5

4

8

4

0

Speed & Cost

Duration

1h 6m

2h 13m

1h 44m

3h 0m

3h 27m

10h 59m

3h 39m

API Cost

$73.44

$61.97

$8.79

$73.92

$28.40

$284.52

$132.06

Code Output

Files Created

388

358

357

376

286

482

389

Classes

170

150

117

150

141

163

153

Lines of Code

6.1K

6.0K

6.0K

6.0K

7.2K

5.7K

4.5K

NCLOC (SonarCloud)

18.6K

17.3K

13.2K

17.3K

16.2K

18.6K

16.4K

Code Quality (PHPMetrics)

Avg Maintainability

79.7

89.7

82.5

89.7

79.6

72.2

72.9

Avg Complexity

4.6

4.5

8.6

4.5

9.2

4.5

4.2

Estimated Bugs

18.1

12.4

4.0

12.4

31.4

14.9

12.2

Violations

36

0

54

25

63

0

0

SonarCloud

Quality GateFailedFailedFailedFailedFailedFailedFailed

Bugs

2

1

3

0

0

1

0

Code Smells

168

102

54

64

153

91

137

Security Hotspots

33

18

0

22

5

47

20

Duplication

2.9%

4.8%

2.8%

7.5%

4.2%

8%

5.9%

ReliabilityCCCAACA

SecurityAAAAAAA

MaintainabilityAAAAAAA

Tech Debt

22.3h

13.4h

12.7h

8.7h

33.3h

14.3h

24.8h

Critical Issues

10

29

6

31

69

47

84

Major Issues

23

27

44

8

53

23

29

Minor Issues

137

47

7

25

31

22

24

Agent Orchestration

Active Agents

31

12

16

29

53

158

34

ModelOpus 4.6Opus 4.6GPT-5.3-codexOpus 4.6GPT-5.3-codexOpus 4.6Opus 4.6

ToolClaude Code v2.1.39Claude Code v2.1.41OpenAI Codex v0.99.0Claude Code v2.1.41OpenAI Codex v0.101.0Claude Code v2.1.80Claude Code v2.1.81

ModeTeam ModeSub-AgentsSub-Agents (experimental)Sub-Agents (tuned)Sub-Agents (quality)Team Mode (advanced prompt)Team Mode (same prompt)

ReasoningThinking: On, Reasoning: MaxThinking: On, Reasoning: MaxReasoning: xhighThinking: On, Reasoning: MaxReasoning: xhighThinking: On, High ReasoningThinking: On, High Reasoning

ParallelismLead + parallel specialistsLead + sub-agentsLead + explorer/worker agentsLead + sub-agents + review agentsLead + explorer/worker agentsLead + controller + QA teammateLead + parallel specialists

* These are hypothetical API costs calculated from list prices. Both sessions ran on fixed-price subscription plans, so no per-token charges actually applied.