Roo Code

🔓 Open

Fork of Cline with custom modes (Architect, Code, Debug), boomerang orchestration, and enhanced context management.

💰 Open-source VS Code extension · vscode, jetbrains

76.1Overall Score

Non-Gameable Scoring

Scores are derived from established benchmarks, adjusted for harness-specific performance across four dimensions: Coding, Reasoning, Tool Use, and Autonomy.

Each dimension starts from public benchmark data and applies harness-specific modifiers based on tool integration, context handling, and orchestration quality. The overall score is a weighted composite that penalizes narrow optimization.

Model	Overall	Coding	Reasoning	Tool Use	Autonomy
Claude Opus 4.6	76.1	74.76	87.25	63.6	88.75
Gemini 3.1 Pro	65.3	81.14	82.49	56.35	75.03
Kimi K2.5	64.6	75.19	75.27	39.07	76.99
GPT-5.4	59.4	47.39	64.69	80.08	79.56
Claude Sonnet 4.6	59.1	68.24	54.65	61.14	87.34
MiniMax-M2.7	57.5	61.57	50.51	62.08	80.03
DeepSeek R1	50.4	68.35	50.51	42.08	56.67
Qwen 3.5	48.9	52.93	50.97	29.04	57.32
Gemini 3 Pro	45.9	44.02	54.94	45.7	58.78
MiMo-V2-Flash	44.0	63.58	46.74	35.86	48.28