Gemma 4 26B A4B

gemma-4-31b-it

Gemma 4 26B A4B is an efficient Mixture-of-Experts (MoE) model from Google DeepMind, combining 26B-class model capacity with approximately 4B active parameters during inference. It delivers strong reasoning, coding, multimodal understanding, and agentic capabilities while offering faster and more efficient inference than larger dense models. With support for text and image inputs, a 256K context window, native function calling, and more than 140 languages, it is well suited for AI agents, coding applications, long-context processing, and high-throughput workloads.

Context Window

262.1K tokens

Maximum Output

32.8K tokens

Release Date

Apr 2, 2026

Modalities

Gemma 4 26B A4B Pricing

Input PriceOutput PriceCache Read
$0.75/M$1/M$0.75/M

Gemma 4 26B A4B API Capabilities

Reasoning

Supported

Tool calling

Supported

Temperature parameter

Supported

Attachments

Supported

Knowledge Base

—

Endpoint Protocols

Completions APIResponses APIMessages API

How to Use Gemma 4 26B A4B via the TokenHub API

Create API key
import OpenAI from "openai"

const client = new OpenAI({
  apiKey: process.env.TOKENHUB_API_KEY,
  baseURL: "https://us-api.tokenhub.com/v1",
})

const result = await client.chat.completions.create({})
console.log(result.choices[0]?.message?.content)

Ready to use Gemma 4 26B A4B?

Use one API key to access Gemma 4 26B A4B and more AI models through TokenHub.

Create API key