Gemma 4 26B A4B IT is an instruction-tuned Mixture-of-Experts (MoE) model from Google DeepMind. It has 25.2B parameters in total, of which only 3.8B are active per token during inference, giving performance close to a 31B model at much lower compute cost. It accepts text, images, and video (up to 60 seconds at 1 fps). Key features include a 256K-token context window, native function calling, an adjustable thinking mode, and structured outputs. It is released under the Apache 2.0 license.
$0.02market price $0.04
$0.005market price $0.01
$0.05market price $0.10
Throughput is how fast the model writes (tokens per second, higher is better). Time to first token is how long before the answer starts (lower is better).
Share of requests served without a server error. Client errors (4xx) do not count.
Works with the OpenAI SDK pointed at our base URL. Every parameter is in the API reference.
Tokens processed per day on NagaAI.
Gemma 4 26B A4B IT is an instruction-tuned Mixture-of-Experts (MoE) model from Google DeepMind. It has 25.2B parameters in total, of which only 3.8B are active per token during inference, giving performance close to a 31B model at much lower compute cost. It accepts text, images, and video (up to 60 seconds at 1 fps).
On NagaAI, Gemma 4 26B A4B costs $0.02 per 1M input tokens and $0.05 per 1M output tokens. You pay only for what you use, with no subscription.
Gemma 4 26B A4B reads up to 262.1k tokens in one request, counting the prompt and the answer together.
Yes. Gemma 4 26B A4B can call your functions as tools and can answer in a JSON schema you give it.
It takes text and images as input and returns text.
Create an API key, copy the model ID from the top of this page and send requests with the OpenAI or Anthropic SDK pointed at NagaAI. The API section on this page has a ready example.
Gemini 4 ArgonGoogleNano Banana 2.1GoogleGemini 3.8 FlashGoogleGemini 3.7 FlashGoogleGemini 3.6 FlashGoogleGemini 3.5 Flash LiteGoogle