2026-09-04
Artificial Analysis
11 related
Artificial Analysis Coding Agent Index: GPT-6 Astra scored 67, roughly equal to Claude Opus 5, Fable 5, and Muse Spark 1.3, but trailing leader Fable 5.1's 70
Fable 5.1 in Claude Code leads the Index with a score of 70. … Various effort levels of the model occupy the Pareto frontier of token efficiency.
Loading articles...