Muse Glimmer 30B muse-glimmer-30b

meta-models · meta-models

unranked — no quality signals yet

params
29.8B
arch
dense
context
128k
license
apache-2.0 open weights
released
Aug 2026
downloads/30d
633.8k

# architecture

attn shared GQA 32q/2kv feed-forward (dense) every parameter, every token out ×52 layers ctx 131,072 29.8B — no routing
layers
52
d_model
6656
heads
32q / 2kv
head dim
128
vocab
202048
family
muse_glimmer_text

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 80.0 ● 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
bench_composite 0.30 80.0

benchmark panel evidence:

complete panelscore (0–100)
aime-2026-v180.0

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-lcr 0.8 / 1 LLM Stats
AIME 2026 0.9 / 1 LLM Stats
Charxiv-r 0.8 / 1 LLM Stats

# where to run

providerquantctx$/M in$/M out$/M cacheprice srctpsuptime
Nvidia 128k via requesty
Phalaunknown 128k $0.30$1.10 phala 98.9%
DeepInfrabf16 128k $0.30$1.20 deepinfra 100.0%
Fireworksunknown 128k $0.35$1.50$0.04 via openrouter 100.0%
Togetherunknown 128k $0.35$1.50$0.04 via openrouter 99.1%

sorted by blended price ((3·input + output) / 4 per 1M) · ✓ = the provider's own catalog confirms the offer · "via …" prices are what the aggregator routing the offer charges, not the provider's own list price

source aliases
llmstats
muse-glimmer-30b
openrouter
meta/muse-glimmer-30b