9B
Parameters
Dense Transformer
AIVory Smart Inference · Gemma
Open WeightGemma 2 9B - nine billion, zero waste.
Google DeepMind's 9B model. The smallest and cheapest text model in the catalogue. Use it for high-volume classification, tagging, and extraction where every fraction of a cent counts.
At a glance
Gemma 2 9B via Smart Inference.
Routed to the cheapest live provider on every request. Same model, lower bill.
8K
Context window
tokens per request
—
Cheapest input price
per 1M tokens via Smart Inference
—%
Cheaper
vs the priciest endpoint for the same model
Pricing
Gemma 2 9B: price comparison.
Our cheapest routed price vs what you'd pay elsewhere.
Prices in USD per 1M input tokens · updated live from the router
Quick start
Use Gemma 2 9B in three lines.
OpenAI-compatible. Swap your base_url, keep everything else.
from openai import OpenAI
client = OpenAI(
base_url="https://smart.aivory.net/v1",
api_key="sk-aivory-...",
)
resp = client.chat.completions.create(
model="gemma2-9b",
messages=[{"role": "user", "content": "Ship it."}],
)
print(resp.choices[0].message.content)import OpenAI from "openai";
const client = new OpenAI({
baseURL: "https://smart.aivory.net/v1",
apiKey: process.env.AIVORY_API_KEY,
});
const resp = await client.chat.completions.create({
model: "gemma2-9b",
messages: [{ role: "user", content: "Ship it." }],
});
console.log(resp.choices[0].message.content);curl https://smart.aivory.net/v1/chat/completions \
-H "Authorization: Bearer $AIVORY_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "gemma2-9b",
"messages": [{"role": "user", "content": "Ship it."}]
}'That's the whole migration. Get early access and try it →
Use cases
What Gemma 2 9B is great at.
-
Millions-Scale Classification
Tag sentiment, detect language, classify intent across your entire dataset. At 9B parameters, per-item cost approaches zero. Process a million items for a few dollars.
-
Entity Extraction Pipelines
Pull names, dates, amounts, and addresses out of text. Gemma 2 is small enough that the per-extraction cost is negligible. Use it as the first pass in a pipeline, then send edge cases to a larger model.
-
Edge and On-Premise Deployment
Fits on a consumer GPU (RTX 3090, 4090) or even some laptop GPUs. Run it locally for data-sensitive workloads. Rent a cheap spot GPU from the <a href='/smart-inference/cheap-gpu/'>GPU Marketplace</a> for dedicated capacity.
FAQ
Common questions.
How cheap is Gemma 2 9B on AIVory?
The cheapest text model in the catalogue. The 9B size means providers serve it on minimal hardware. Check the live table above for exact per-token rates.
When is 9B not enough?
For anything requiring reasoning, nuance, or creative output. Gemma 2 9B works well for classification, extraction, and simple transformations. If you need multi-step logic or long, coherent text, step up to Gemma 4 31B or a 70B model.
Why is the context window only 8K?
Gemma 2 was trained before the 128K context trend. 8K tokens is enough for most classification and extraction tasks (typical inputs are under 1K tokens). For long documents, use Mistral Nemo 12B (128K context, similar price range).
Can I run Gemma 2 9B on my own GPU?
Yes. It runs on any GPU with 16GB+ VRAM at full precision, or 8GB+ at 4-bit quantization. Rent one from the GPU Marketplace or use your own hardware.
Start using Gemma 2 9B. Today.
Swap your base_url, keep shipping.
The router is live. Credits from $10, no subscription, no commitment.