Lyria 3 Pro Preview has the larger context window (1,048,576 vs 163,840 tokens, 6.4×). Llama Guard 4 12B is open-weights; the other is not.

Lyria 3 Pro Preview vs Llama Guard 4 12B

Live catalog fields. Best-in-row is highlighted.

Lyria 3 Pro Preview

Google · google__lyria-3-pro-preview

Llama Guard 4 12Bopen

Meta · meta-llama__llama-guard-4-12b

Identity

Fieldgoogle__lyria-3-pro-previewmeta-llama__llama-guard-4-12b
ProviderGoogleMeta
Sluggoogle__lyria-3-pro-previewmeta-llama__llama-guard-4-12b
Statuslivelive
Open weightsNoYes
License——
HuggingFace—meta-llama/Llama-Guard-4-12B

Quality

Fieldgoogle__lyria-3-pro-previewmeta-llama__llama-guard-4-12b
Quality index (0–100)——

Cost

Fieldgoogle__lyria-3-pro-previewmeta-llama__llama-guard-4-12b
Input $ / 1M tokens$0.00$0.18
Cached input $ / 1M——
Output $ / 1M tokens$0.00$0.18

Context

Fieldgoogle__lyria-3-pro-previewmeta-llama__llama-guard-4-12b
Max context tokens1,048,576163,840
Max output tokens65,53616,384

Modalities

Fieldgoogle__lyria-3-pro-previewmeta-llama__llama-guard-4-12b
Inputtext, imageimage, text
Outputtext, audiotext