Google DeepMind's experimental diffusion-based member of the Gemma 4 open model family. Unlike autoregressive models that generate text one token at a time, DiffusionGemma denoises a canvas of placeholder tokens to produce up to 256 tokens in parallel, finalizing output in one block. A Mixture-of-Experts model with 26B total parameters and 3.8B active per inference, delivering roughly 4x the throughput of similarly sized autoregressive Gemma models on local hardware. Excels at non-linear tasks like in-line editing, molecular sequencing, mathematical graphing, and self-correcting puzzles.
Cheapest
Google AI Studio
Free / Free
Google AI Studio, Hugging Face Inference
Sorted by total cost (input + output per 1M tokens). Select a row to view provider details.
| Provider | Pricing (per 1M) | Rate limits | Regions | Health | Latency |
|---|---|---|---|---|---|
In: FreeOut: Free | 15 RPM / 500K TPM | us-east-1eu-west-1global | Healthy | Unknown | |
In: $0.15Out: $0.15 | 300 RPM / 500K TPM | us-east-1eu-west-1 | Healthy | Unknown |
Use this model via Google AI Studio with an OpenAI-compatible SDK.
import OpenAI from "openai";
const client = new OpenAI({
baseURL: "https://generativelanguage.googleapis.com/v1beta/openai",
apiKey: process.env.GOOGLE_AI_STUDIO_API_KEY,
});
const response = await client.chat.completions.create({
model: "diffusiongemma-26b-a4b-it",
messages: [
{ role: "user", content: "Hello!" }
],
});
console.log(response.choices[0].message.content);Using Google AI Studio API · OpenAI-compatible SDK
Every price recorded for this model, per provider. Prices are per 1M tokens in USD.
No price changes recorded since Jun 11, 2026.
| Date | Provider | Input | Output |
|---|---|---|---|
| Jun 11, 2026Listed · Current | Hugging Face Inference | $0.15 | $0.15 |
| Jun 11, 2026Listed · Current | Google AI Studio | $0 | $0 |