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)

TechmemeSeptember 4, 2026en

Artificial Analysis : 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.

This is a short summary published by AI Global Wire. The full article is owned and hosted by Techmeme — open it there to read it in full.

• OpenAI
• Anthropic
• Agenter