Gemma 4 26B A4B IT

Text

Gemma 4 26B A4B IT

Efficient Mixture-of-Experts model optimized for strong performance at lower cost.

Efficient MoE Design. ~26B total parameters with ~4B active per token. Delivers near large-model quality at significantly lower compute cost.
Instruction Tuned. Built for chat, structured outputs, and general assistant use cases.
Cost-Performance Sweet Spot. Much cheaper to run than dense models of similar quality, making it practical for production workloads.
Apache 2.0 License. Fully open and permissive for commercial deployment.

Why pick it for Norman AI?

Gemma 4 26B A4B IT is a great “default at scale” model - good quality, lower cost, and easy to run. Ideal when you need to serve many requests without paying for full dense models.

messages = [
    {"role": "system", "content": "You are a helpful AI assistant."},
    {"role": "user", "content": "Can you provide ways to eat combinations of bananas and dragonfruits?"},
    {"role": "assistant",
     "content": "Sure! Here are some ways to eat bananas and dragonfruits together"},
    {"role": "user", "content": "What about solving an 2x + 3 = 7 equation?"},
]

response = await norman.invoke(
    {
        "model_name": "gemma-4-26b-a4b-it",
        "inputs": [
            {
                "display_title": "Prompt",
                "data": messages
            }
        ]
    }
)

messages = [
    {"role": "system", "content": "You are a helpful AI assistant."},
    {"role": "user", "content": "Can you provide ways to eat combinations of bananas and dragonfruits?"},
    {"role": "assistant",
     "content": "Sure! Here are some ways to eat bananas and dragonfruits together"},
    {"role": "user", "content": "What about solving an 2x + 3 = 7 equation?"},
]

response = await norman.invoke(
    {
        "model_name": "gemma-4-26b-a4b-it",
        "inputs": [
            {
                "display_title": "Prompt",
                "data": messages
            }
        ]
    }
)

messages = [
    {"role": "system", "content": "You are a helpful AI assistant."},
    {"role": "user", "content": "Can you provide ways to eat combinations of bananas and dragonfruits?"},
    {"role": "assistant",
     "content": "Sure! Here are some ways to eat bananas and dragonfruits together"},
    {"role": "user", "content": "What about solving an 2x + 3 = 7 equation?"},
]

response = await norman.invoke(
    {
        "model_name": "gemma-4-26b-a4b-it",
        "inputs": [
            {
                "display_title": "Prompt",
                "data": messages
            }
        ]
    }
)

View Docs