Skip to content
AI Atlas
Familyroot · Llama

Llama 4

All 5 models in the terminal →

first seen 11 Sept 2026

family_01M2A55V2Y87QSWA1SXY37QN3Z

Models
5
First release
2 Apr 2025
Latest release
5 Apr 2025
Parameters
108.6B
Artifacts
0
Providers
1

Releases

Family releases by statusActiveApr 25Apr 25Apr 25Llama 4 Scout — 2 Apr 2025Llama 4 Maverick — 5 Apr 2025

2 dated releases · dot size = best benchmark rank of the member · 3 members without a sourced release date not shown

Members 5

Family members
ModelParamsContextOpennessLicenceReleasedStatusBest rankCompare
Llama 4 MaverickMeta AI—1.05MOpen weights—5 Apr 2025Active#17 Aider polyglot — well-
Llama 4 ScoutMeta AI108.6B1.31MOpen weightsOther2 Apr 2025Active#122 SciCode
Llama 4 Maverick InstructMeta AI—————Active#39 SWE-bench Verified
Llama 4 Scout InstructMeta AI—————Active#41 SWE-bench Verified
Llama-4-Scout-17B-16EMeta AI——Open weights——Active—

Lineage among members 1 relations

Open in Graph →
Lineage of the Llama 4 familyLlama 4 Scout fine-tuned from Llama-4-Scout-17B-16ELlama 4 Scout108.6BLlama 4 ScoutLlama-4-Scout-17B-16ELlama-4-Scout-17B-16E

Benchmark progress

MMMU-Probest #100

Best rank of the family on MMMU-Pro#150#100#50#0Apr 25Apr 25Apr 25Apr 25Apr 25Apr 25

SciCodebest #108

Best rank of the family on SciCode#150#100#50#0Apr 25Apr 25Apr 25Apr 25Apr 25Apr 25

IFBenchbest #181

Best rank of the family on IFBench#250#200#150#100#50#0Apr 25Apr 25Apr 25Apr 25Apr 25Apr 25

Terminal-Benchbest #184

Best rank of the family on Terminal-Bench#300#200#100#0Apr 25Apr 25Apr 25Apr 25Apr 25Apr 25

GPQA Diamondbest #245

Best rank of the family on GPQA Diamond#400#300#200#100#0Apr 25Apr 25Apr 25Apr 25Apr 25Apr 25

Artificial Analysis Intelligence Indexbest #249

Best rank of the family on Artificial Analysis Intelligence Index#500#400#300#200#100#0Apr 25Apr 25Apr 25Apr 25Apr 25Apr 25

Best rank (primary comparability group, current leaderboard) reached by any member released up to each date — lower is better, axis inverted. Ranks are today's ranks, not the ranks at release time.

Aggregates are computed from the members' sourced claims; members without a release date or parameter count are left out of the range and timeline rather than guessed. Methodology →