Motif 3 motif-3

Motif-Technologies · motif-technologies

unranked — no quality signals yet

params
314.8B
arch
moe
license
mit open weights
released
Aug 2026
train compute
1.1e+24 FLOP
downloads/30d
5.6k

# architecture

attn shared GQA 80q/16kv router expert ×384 expert out ×53 layers 314.8B pool
layers
53
d_model
4096
heads
80q / 16kv
head dim
192
experts
top-0 of 384 + 1 shared
vocab
220160
family
Motif

signal path of one layer, generated from the registry's structured fields — dimension lines quote the model's real numbers; an MoE trace forks at the router, a dense trace runs straight through. Geometry fields come from the repo's config.json.

# why ranked

Overall: provisional — score 93.8 ● low — a single signal; treat with caution

provisional: too few independent signals for a numbered position — the score is shown, but this model sorts after every ranked model.

signalweightinput (0–100)
aa_intelligence 0.70 93.8
bench_composite 0.30

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

Coding: provisional — score 89.5 ● low — a single signal; treat with caution

provisional: too few independent signals for a numbered position — the score is shown, but this model sorts after every ranked model.

signalweightinput (0–100)
aa_coding 0.30 89.5
aider_polyglot 0.30
swe_bench_verified 0.40

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

full methodology

# trend

OGM score · last 26 days
OGM score over 26 days

# benchmarks

benchmarkscoresourcedate
AA Coding Index 63.5 Artificial Analysis
AA Intelligence Index 38.0 Artificial Analysis

# where to run

no known API providers — download the weights from Hugging Face and run them yourself.

source aliases
epoch
Motif-3