Llama 4 Maverick llama-4-maverick

meta-llama · llama flagship

#21 overall #28 coding

params
400B (A17B)
arch
moe
context
1024k
license
llama4 open weights gated
released
Apr 2025
train compute
2.2e+24 FLOP
downloads/30d
51.9k

# why ranked

Overall: #21 — score 14.3 ● low — a single signal; treat with caution

signalweightinput (0–100)
aa_intelligence 0.50 14.3
arena_elo 0.30
bench_composite 0.20

missing signals are dropped and the remaining weights renormalized — never imputed.

Coding: #28 — score 22.4 ● medium — one signal missing, weights renormalized

signalweightinput (0–100)
aa_coding 0.25 35.7
aider_polyglot 0.25 9.1
arena_elo_code 0.15
swe_bench_verified 0.35

missing signals are dropped and the remaining weights renormalized — never imputed.

full methodology

# trend

OGM score · last 2 days
OGM score over 2 days
overall rank (up is better)
Overall rank over 2 days

# benchmarks

benchmarkscoresourcedate
AA Coding Index 16.3 Artificial Analysis
AA Intelligence Index 14.3 Artificial Analysis
AA Math Index 19.3 Artificial Analysis
Aider Polyglot 15.6 Aider Apr 2025

# where to run

providerquantctx$/M in$/M out$/M blendtpsuptime
DeepInfrafp8 1024k $0.15$0.60$0.26 100.0%
DigitalOceanunknown 125k $0.25$0.87$0.41 100.0%
Novitafp8 1024k $0.27$0.85$0.42 99.9%
Parasailfp8 512k $0.35$1.00$0.51 100.0%
Googleunknown 512k $0.35$1.15$0.55 99.9%

blended = (3·input + output) / 4 per 1M tokens · cheapest first

source aliases
aa
llama-4-maverick
aider
Llama 4 Maverick
arena
Llama-4-Maverick-17B-128E-Instruct
epoch
Llama 4 Maverick
openrouter
meta-llama/llama-4-maverick