AGIdex
agidex/Technologies/Llama 4 Maverick
AI MODEL

Llama 4 Maverick

OPEN WEIGHTS
RELEASED 2025-04

Meta's flagship open-weight model. 400B/17B active MoE architecture, 1M token context.

Benchmark coverage · per sub-capability

Each score is anchored to human baseline (100). Source URLs link to the original benchmark, leaderboard, or release note.

04ReasoningCOGNITIVE
Sub-capabilityQualitySourceScore vs. baselineScore
Common Sense Reasoningvendor reportedai.meta.com/llama-4
66%
01CommunicationCOGNITIVE
Sub-capabilityQualitySourceScore vs. baselineScore
Language Understandingvendor reportedai.meta.com/llama-4
92%