Skip to content
Bencher

TestsBenchmarks · Agentic

Agentic IF Index (Meta internal)

Meta-internal composite of instruction following in agentic tasks.

% solvedHigher is betterThe test's websiteOfficial page ↗

The best 2 · 2 models tested

Top 2 · best setting per model · 2 models, 2 results (0 independent)

Measured by independent testersIndependent Reported by the makerVendor-reported
Agentic IF Index (Meta internal) leaderboard
Thinking levelSetting
1Muse Spark 1.3Meta (Meta Superintelligence Labs, Muse)57.8%Maxmax reasoningMaker's own figureVendor-reportedMeta ↗—2 Sep 2026
2Muse Spark 1.2Meta (Meta Superintelligence Labs, Muse)46.2%Extra highMaker's own figureVendor-reportedMeta ↗—2 Sep 2026