Phi 4 phi-4

microsoft · microsoft

unranked — no quality signals yet

params
14.7B
arch
dense
context
16k
license
mit open weights
released
Dec 2024
downloads/30d
360.1k

# architecture

attn shared GQA 40q/10kv feed-forward (dense) every parameter, every token out ×40 layers ctx 16,384 14.7B — no routing
layers
40
d_model
5120
heads
40q / 10kv
head dim
128
vocab
100352
family
phi3

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.

# trend

OGM score · last 88 days
OGM score over 88 days
overall rank (up is better)
Overall rank over 88 days

methodology changed during this history window; score movement across that boundary is not model movement. See methodology v6.

# benchmarks

benchmarkscoresourcedate
Arena-hard 0.8 / 1 LLM Stats —

# where to run

providerquantctx$/M in$/M out$/M cacheprice srctpsuptime✓
Featherless — ——— via hfrouter — —
DeepInfrabf16 16k $0.07$0.14— deepinfra — 100.0% ✓
DeepInfrabf16 16k $0.07$0.14— deepinfra — 99.8% ✓
Azure 125k $0.12$0.50— via modelsdev — —

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
aa
phi-4
llmstats
phi-4
openrouter
microsoft/phi-4