benchgap
xAI · model

Grok 4.20 Multi-agent benchmark scores

As of 2026-11-18, Grok 4.20 Multi-agent (xAI) has measured scores on 1 benchmark and estimated scores on 12 more.

BenchmarkScoreSource
FailureSensorIQ68.0%estimated ± 1.7 pp, high confidence
Chess Puzzles48.3%estimated ± 12.3 pp, low confidence
Mystery Game Puzzles45.3%estimated ± 14.6 pp, low confidence
NYT Connections89.6%measured
SpireClimb7.0%estimated ± 8.8 pp, medium confidence
StealthMark58.3%estimated ± 10.7 pp, low confidence
SkillsBench official leaderboard (BenchFlow)55.5%estimated ± 6.1 pp, medium confidence
BinaryAudit35.8%estimated ± 12.1 pp, low confidence
OTelBench22.5%estimated ± 6.6 pp, medium confidence
FACTS Multimodal50.2%estimated ± 4.0 pp, high confidence
BlenderWorld (text-only)81.8%estimated ± 4.3 pp, medium confidence
Cookie-Bench89.2%estimated ± 4.8 pp, medium confidence
ExploitBench Internal Port15.3%estimated ± 13.8 pp, low confidence