Muse Spark 1.2 ranks #5 on GDPval-AA v2, our benchmark of agentic real-world…
- Source
- x.com
- Date
Muse Spark 1.2 ranks #5 on GDPval-AA v2, our benchmark of agentic real-world knowledge work, at 1631 Elo - behind e.g. Claude Opus 5 (max, 1852), GPT-5.6 Sol (max, 1730), and Kimi K3 (1685), and ahead of e.g. Claude Opus 4.8 (max, 1588). Muse Spark 1.1 scored 1371 at its launch last month
Context
Meta's Muse Spark 1.2 placed fifth in Artificial Analysis's August 5 GDPval-AA v2 results, with an Elo rating of 1631. Version 1.1 had scored 1371 at its launch.
GDPval evaluates work products such as documents and spreadsheets; its evaluation framework derives ratings from comparisons between outputs. Elo is a relative rating, not a percentage of tasks completed or a measure of employee productivity. The dated result supports improved performance on evaluated knowledge-work tasks, with the rank applying to that version of the .
- eval — A repeatable test for AI quality — a set of tasks plus scoring — used the way software teams use test suites, because model output is too variable to judge by eyeballing.
- benchmark — A standard public test set for comparing AI models — the shared scoreboards behind every "model X beats model Y" claim.
Checking sign-in…
Loading comments…