Preseason
MatchesRankingsPrompts
GitHub
Preseason
MatchesRankingsPromptsMethodologyContact

© 2026 Preseason. All rights reserved.

PrivacyTerms
@betocmn
Agentic IDE / ADEs
Methodology

Devin vs Codex CLI

DevinDEDevinvsCodex CLICOCodex CLI
DevinCodex CLI
50%
50%
Insufficient data
This matchup has 8 decisive cases (minimum 30 required for publication).

Statistics

MetricValue
Devin wins4
Codex CLI wins4
Abstains (no tool)880
Other tool chosen2579
Decisive cases8
Devin win rate (unweighted)50.0%
95% CI21.5% - 78.5%
Devin win rate (weighted)50.0%

Comments

Devin

No comments yet

Verified critics can leave comments here.

Codex CLI

No comments yet

Verified critics can leave comments here.

Per-model breakdown

ModelTierDevinCodex CLINoneOtherA rate
DeepSeek V4 ProFrontier20844100%
GPT 5.4Frontier0231950%
GPT 5.4 MiniMid0281740%
Gemini 3.5 FlashSmall10838100%
MiMo V2.5 ProFrontier101043100%
Claude Haiku 4.5Small008290n/a
Claude Opus 4.6Frontier0020112n/a
Claude Opus 4.8Frontier00846n/a
Claude Sonnet 4.6Frontier0073113n/a
DeepSeek R1 0528Frontier0065118n/a
DeepSeek V3.2Mid005278n/a
DeepSeek V4 FlashMid001240n/a
Devstral 2 2512Mid0064106n/a
Gemini 2.5 FlashSmall005973n/a
Gemini 2.5 ProFrontier0022159n/a
GLM 5 TurboFrontier002130n/a
GLM 5.2Frontier00054n/a
GPT 5.3 CodexFrontier0027156n/a
GPT 5.5Frontier00252n/a
Kimi K2.5Frontier005365n/a
Kimi K2.7 CodeFrontier00647n/a
Llama 4 MaverickFrontier0066118n/a
Llama 4 ScoutSmall0053120n/a
MiMo V2 ProFrontier0030102n/a
MiniMax M2.7Frontier0025104n/a
MiniMax M3Frontier00450n/a
Mistral Small 4Mid006991n/a
Qwen3 Coder NextMid0021161n/a

Per-prompt breakdown

PromptTierDevinCodex CLINoneOtherA rate
ai-support-agent-platformAdvanced14051620%
ai-engineering-workflowAdvanced30092100%
ai-engineering-workflowBeginner000114n/a
ai-engineering-workflowIntermediate000114n/a
ai-revenue-ops-copilotIntermediate002516n/a
ai-revenue-ops-copilotBeginner00281252n/a
ai-revenue-ops-copilotAdvanced008495n/a
ai-support-agent-platformBeginner00339197n/a
ai-support-agent-platformIntermediate00250283n/a