Coding Agent Index
An evaluation suite specifically measuring model performance on autonomous coding tasks, requiring the model to complete tasks independently within loops of multi-step tool calls, file reading and writing, test execution, and error fixing, scored by task pass rate or quality standards.
Core Facts
Timeline (last 90 days)
Gemini 4 Neon ranks first on the AA Coding Agent Index with 59 points.
The Coding Agent Index requires models to complete tasks independently in loops of multi-step tool calls, file reading and writing, test running, and error fixing, scoring by task pass rate or quality standards.
All Facts (2)
Gemini 4 Neon ranks first on the AA Coding Agent Index with 59 points.
70%UnverifiedThe Coding Agent Index requires models to complete tasks independently in loops of multi-step tool calls, file reading and writing, test running, and error fixing, scoring by task pass rate or quality standards.
50%