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 (Artificial Analysis)

The Artificial Analysis Coding Agent Index shows GPT-6 Astra scored 67, placing it roughly equal to Claude Opus 5, Fable 5, and Muse Spark 1.3. However, GPT-6 Astra trails the current leader, Fable 5.1, which achieved a score of 70. Fable 5.1, specifically in Claude Code, leads the Index, demonstrating its position on the Pareto frontier of token efficiency among various model effort levels.