Megatron MarketsWholesale
Sign inContact Us
Models Model details

Google

Google: Gemma 4 26B A4B

google/gemma-4-26b-a4b-it

Enquire about this model Request a quote

Gemma 4 26B A4B IT is an instruction-tuned Mixture-of-Experts (MoE) model from Google DeepMind. Despite 25.2B total parameters, only 3.8B activate per token during inference — delivering near-31B quality at a fraction of the compute cost. Supports multimodal input including text, images, and video (up to 60s at 1fps). Features a 256K token context window, native function calling, configurable thinking/reasoning mode, and structured output support. Released under Apache 2.0.

Input
$0.0675

/ 1M tokens

Output
$0.225

/ 1M tokens

Context window
262.14K

tokens

Limits

Context

262,144

tokens

Max output

235,929

90% of the context window

Parameters

25,805,936,206

Model facts

Author

Google

Tokenizer

Gemma

Modality

Text + Image + Video -> Text

License

Apache-2.0

Created

Apr 3, 2026

Hugging Face

google/gemma-4-26B-A4B-it

Default generation parameters

temperature

1

top_k

64

top_p

0.95

Publication status

Hugging Face: verifiedLicense: publishedParameter count: published

Modalities

Input · ImageInput · TextInput · VideoOutput · Text

Generation controls

Frequency penaltyInclude reasoningLogit biasLogprobsMax tokensMin pPresence penaltyReasoningRepetition penaltyResponse formatSeedStopStructured outputsTemperatureTool choiceToolsTop kTop logprobsTop p

Capability scores

2.9
Agentic index

Better than 23% of models compared

39.3
Coding index

Better than 33% of models compared

16.7
Intelligence index

Better than 41% of models compared