Gemma 3 27B IT
google/gemma-3-27b-it
Google· 98K ctx· gemma
open weights quant: unknown
spread (output, p90 ÷ p10)
4.3×
$0.160 – $0.760 · 8 providers · max÷min 4.8×
8 providers priced · 8 offerings
Cheapest / median / spread use providers that serve the model on their own capacity: first-party lab, hyperscaler, inference host.
Blended $/M assumptions
blended = 90.9% input × [(1 − 70% cache) × input + 70% cache × cache-read] + 9.1% × output + cache writes (30% of uncached input) where a write price is published
A provider that publishes no cache price is blended with cache share 0 (all input at the input price) and labelled “no cache price published” — never penalised against identical in/out prices.
| Provider | Input | Output | Cache rd | Blended |
|---|---|---|---|---|
Eden AI router / aggregator | $0.0800 | $0.160 | — | $0.0873no cache price published |
Kilo Gateway router / aggregator | $0.0800 | $0.160 | — | $0.0873no cache price published |
Deep Infra inference host | $0.0800 | $0.160 | — | $0.0873no cache price published |
Hugging Face inference host | $0.0800 | $0.160 | — | $0.0873no cache price published |
NovitaAI inference host | $0.119 | $0.200 | — | $0.126no cache price published |
Venice AI inference host | $0.120 | $0.200 | — | $0.127no cache price published |
Cortecs router / aggregator | $0.0990 | $0.299 | — | $0.117no cache price published |
NanoGPT router / aggregator | $0.299 | $0.299 | $0.150 | $0.204 |
Nebius Token Factory inference host | $0.100 | $0.300 | $0.0100 | $0.0711 |
Amazon Bedrock hyperscaler | $0.230 | $0.380 | — | $0.244no cache price published |
OpenRouter router / aggregator | $0.0800 | $0.450 | $0.0400 | $0.0882 |
Merge Gateway router / aggregator | $0.0800 | $0.450 | $0.0400 | $0.0882 |
GreenPT inference host | $0.342 | $0.684 | — | $0.373no cache price published |
STACKIT inference host | $0.530 | $0.760 | — | $0.551no cache price published |
Data: models.dev (MIT), snapshot 2026-09-20. Prices are vendor-published; verify before purchase. Methodology.