#14

Copilot Workspace

🔒 Closed

GitHub-integrated AI coding environment with Codespaces, PR-native workflow, and repository-wide context.

💰 Included with GitHub subscription · web, vscode

60.7Overall 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	60.7	74.15	71.91	39.88	51.27
Gemini 3.1 Pro	51.5	80.55	69.01	34.36	42.12
Claude Sonnet 4.6	44.4	66.86	44.47	37.18	49.18
GPT-5.4	44.0	46.87	53.15	49.36	45.63
Gemini 3 Pro	34.9	43.64	45.29	27	32.54
Gemini 3 Flash	31.6	50.58	44.3	19.39	23.16
GPT-5.2	30.6	49.41	51.75	23.84	28.15
Grok 3	30.1	68.38	31.75	8.08	13.05
Claude Sonnet 4.5	28.8	54.47	36.78	18.77	34.16
GPT-5 (high)	28.8	53.26	41.67	18.82	30.31