Can I run Gemma 4 26B A4B locally?

Yes — the weights are open under Apache 2.0 and available on Hugging Face at google/gemma-4-26B-A4B-it, so you can download and run it yourself, though the 25.2B total parameter count means you will need a capable machine.

What does it cost via API, and how long is the context?

Pricing is $0.09 per million input tokens and $0.30 per million output tokens, with a 256K token context window.

How does the 26B A4B differ from the 31B Dense in the Gemma 4 family?

The 26B A4B is a Mixture-of-Experts model that activates only 3.8B parameters per token, making it cheaper to run than the 31B Dense while targeting similar output quality; the 31B Dense activates all its parameters every time and does not support video input.

▲

53.5%

this week

Gemma 4 26B A4B

8770°

#llm #google #vision #video #reasoning #agents

Run this model

Offers / Ads

OpenRouter Terminal / API

Pay per token

Chat with this model in the browser or call it from your code with one API key.

Google | 262,144 token context | $0.09/M input, $0.30/M output. 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.

Guides for this model

How to Install Ollama and Run a Local AI (Windows, Mac, Linux)

Ollama vs LM Studio (2026): which local AI app to use, and which top models each one runs

What is a GGUF file? Quantized models explained (Q4, Q8, fp8, NF4)

How to choose an AI model in October 2026: the closed source news, then the open weights worth downloading

Best local LLM by use case (October 2026): the right open model for every job

How much VRAM do I actually need for local AI?

All guides

Captain's notes

Written by Captain, the house editor

Notes from 2026-10-13

Gemma 4 26B A4B is Google DeepMind's instruction-tuned Mixture-of-Experts model that activates only 3.8B of its 25.2B parameters per token, so you get near-31B-class output at a fraction of the compute cost. It handles text, images, and video up to 60 seconds, with a 256K-token context window and a configurable reasoning mode. Apache 2.0 weights make it a serious option for anyone who wants a capable multimodal model they can actually run and modify.

Known for

MoE efficiency: 3.8B active parameters per token, near-31B quality
Multimodal: accepts text, images, and video up to 60 seconds
256K token context window for long-document work
Native function calling and structured output for agentic pipelines
Apache 2.0 licence — use it, modify it, deploy it commercially

Watch out for

Video input is capped at 1fps and 60 seconds; do not expect rich temporal understanding.
Audio is not supported on this size — only on the E2B, E4B, and 12B variants.
At 25.2B total parameters, local deployment requires a well-specced machine despite the MoE savings.

Best for

Prompts to try

Debug a tricky async bug

Here is a Python async function that occasionally raises a RuntimeError: 'Event loop is closed' when called from a FastAPI background task. Walk through exactly why this happens, then rewrite the function so it is safe to call in that context. Show the corrected code and explain each change.

Reason through a scheduling puzzle

Five engineers must each review two of seven pull requests. No engineer reviews their own PR. PRs 3 and 7 must be reviewed by engineers who have not reviewed any of PRs 1, 2, or 4. Assign reviewers to all seven PRs satisfying these constraints, and show your reasoning step by step.

Analyse an attached image

Attach a screenshot of a dashboard, chart, or UI. Then ask: Describe every data point and visual element you can see, identify any anomalies or trends, and suggest one concrete change that would make this easier to read at a glance.

Compare with

Gemini 3.8 Flash
API-only model
1,048,576 token context · $0.83/M input, $4.12/M output
Gemini 3 Flash Preview
API-only model
1,048,576 token context · $0.50/M input, $3.00/M output
Gemini 2.5 Flash Lite
API-only model
1,048,576 token context · $0.10/M input, $0.40/M output
Gemini 3.1 Flash Lite
API-only model
1,048,576 token context · $0.25/M input, $1.50/M output

More text models from google

Questions people ask

Can I run Gemma 4 26B A4B locally?: Yes — the weights are open under Apache 2.0 and available on Hugging Face at google/gemma-4-26B-A4B-it, so you can download and run it yourself, though the 25.2B total parameter count means you will need a capable machine.
What does it cost via API, and how long is the context?: Pricing is $0.09 per million input tokens and $0.30 per million output tokens, with a 256K token context window.
How does the 26B A4B differ from the 31B Dense in the Gemma 4 family?: The 26B A4B is a Mixture-of-Experts model that activates only 3.8B parameters per token, making it cheaper to run than the 31B Dense while targeting similar output quality; the 31B Dense activates all its parameters every time and does not support video input.

Captain's verdict
A rare combination: open weights, genuine multimodal capability, and MoE efficiency that keeps inference costs honest.

Model card on Hugging Face

Captain writes these notes from the model's own description and model card. Prices and context sizes are the provider's at the time of writing and can change.

Ranking badge

Updated Oct 22, 2026

Made this model? Show this badge on your page.

Paste it into your Hugging Face card or Civitai description.

Change colors / personalize / get code

Also ranked: #46 Hottest Text / LLM model · #46 Hottest Text / LLM text model

Not this creator?

Get your own badge

Gemma 4 26B A4B Popularity Trend

Heat score since 3 Apr 2026

Heat score

8,770°

This week

▲

53.5%

8,713° then, 8,770° now

This month

▲

607%

8,512° then, 8,770° now

Past year

n/a

Wasn't released/tracked

7,000°

Floor

The scale runs from 0° to 9,999°. Everything below 7,000° is kept for models no source reports on yet, which carry a stand-in number there so lists stay complete. 7,000° is where measured data begins: a model with real data but almost no downloads sits at the floor, and the measurements fill the degrees above it. The floor says nothing about the model; only the degrees above it do, and they are what separates one model from another.

+1,489°

OpenRouter tokens

919,242,238,645 tokens this model processed through OpenRouter in the last 30 days, prompts and answers together. Every 10,000,000 tokens count as one download in the download term, 91,924 in all; the last day's tokens join the adoption rate. Source: OpenRouter (openrouter.ai/rankings).

+281°

Adoption rate

About 5,610 downloads a day, measured between two readings of the sources, never guessed from a lifetime average. Speed weighs a quarter of the download term, so a fresh surge shows without letting a one-day spike outrank years of use.

Show detailed breakdown

Every number on this site can be checked. This model's heat score is 8,770°, and here it is built one ingredient at a time, with the running total after each: the floor, the degrees each source adds, the bonus for the kind of model and what the age of its weights takes away. The colors are the bands of the chart above. The glossary at the end defines every term.

1
Start at the floor.
Why 7,000° and not zero: the scale runs from 0° to 9,999°, and the part below 7,000° is reserved for models no source reports on yet, which carry a stand-in number there so the lists stay complete and never outrank a model with data. 7,000° is where measured data begins. A model with real data but almost no downloads sits here; everything that follows is measurement, and the room above the floor holds ten tenfold steps in downloads at 300° each. The floor is the same for every model, so it decides nothing; the degrees above it do.
7,000°
running total
2
Add the downloads.
91,924 download-equivalents in all, from OpenRouter tokens. Ten times the downloads adds the same 300°, so size matters but never runs away with the score.
+1,489°
8,489°
running total
3
Add the adoption rate.
About 5,610 downloads a day right now, measured between two readings, at a quarter of the weight.
+281°
8,770°
running total
=
The heat score.
Where this model stands on every list of this site. Nothing else adjusts it.
8,770°

Glossary: the terms used here

Heat score: One number for how much a model is used and wanted right now, from 0° to 9,999°. Every list on this site that says hottest is sorted by it and nothing else.
Floor: The 7,000° every model with real data starts from. The scale runs from 0° to 9,999°; the part below the floor is reserved for stand-ins (models without data), so that a model with data always ranks above one without. The floor is the same for everyone and decides nothing; the degrees above it are the measurement.
Download-equivalents: Everything the score treats as a download, added up: downloads themselves, each torrent seeder as 1,000, each AI Horde image and each run on a generation site as one.
Log scale: Counting by tens, not by ones: going from 1,000 to 10,000 downloads adds the same degrees as going from 10,000 to 100,000. It keeps giants from crowding out everything else.
Likes: Thumbs up and likes from every source, worth half a download each on the log scale.
Adoption rate: How many downloads a day a model gets right now, from the difference between two readings of a source. A quarter of the weight of the download term.
Measured and estimated: Measured means two readings of a source exist and the difference is real. Estimated means only today's total is known and the past is spread over the model's life; the site says which one it is showing.
Seeder: Someone who keeps a copy of the weights online for others to download over a torrent, or a hosted mirror doing the same. Each one counts as 1,000 downloads.
AI Horde: A volunteer network that generates images for free on donated computers. The number of images it has made with a model is public, and each counts as one download.
Generation sites: Sites that host their own copies of a model and let people generate with it, such as PixAI, SeaArt and Yodayo. Their downloads count like Civitai's and each generation run as one download.
OpenRouter tokens: A text model served through an API, such as Claude or GPT, has no weights to download, so its usage is measured in tokens: the pieces of text it read and wrote through OpenRouter, a gateway many apps use, over the last 30 days. Every 10,000,000 tokens count as one download. Source: OpenRouter (openrouter.ai/rankings), CC BY 4.0.
Full model (checkpoint): The weights themselves, the file that generates on its own. It gets a bonus because add-ons depend on it.
LoRA and embedding: Small add-on files that steer someone else's weights toward a style, a character or a concept. They cannot run alone.
Family and weights: The base a model is built on, such as SD 1.5, SDXL, Flux or Wan. Every model inherits the release date of its family's weights.
Age: Degrees taken off for how long ago the family's weights came out: 150° a year, up to 750°. It keeps old downloads that had years to pile up from freezing the top of the list.
Family boost: A small, bounded nudge for LoRAs of a family this site is pushing forward right now, always listed in the breakdown.
Stand-in score: A stable number below the floor given to a model no source reports on yet, so lists stay complete. It never ranks above a model with data.
New! (grace period): A model less than 7 days old shows New! instead of a score and holds no place in the rankings. Its score is computed from day one, but a model one day old grows faster than anything, so without the wait the week's #1 would always be whatever came out yesterday rather than what held up for a week.
Trend: How a model's growth over a period compares with the typical model's growth over the same period. Faster than typical reads as a rise, slower as a fall.
Typical model: The middle of the catalog: half the models with a figure grew faster over the period, half slower.

How this model's popularity has moved against the rest of the catalog: its growth in downloads, likes and usage over each period, measured against the typical model's growth over the same period, with its heat score at the time. Above the typical growth reads as a rise, below it as a fall. The past is reconstructed from today's totals, the model's first day and its measured download speed where two readings exist; a model that did not exist at the start of a period shows n/a.

Similar Models

Meta Llama 3 70B Instruct

TEXT / LLM

AI Models for Image, Video & Text Generation

Gemma 4 26B A4B is a text model for the Text / LLM family listed on ModelTrends.ai, a read-only catalog of open source AI image, video and text models. Compare it with other Text / LLM models and check its heat score to see how it is trending.

More Text / LLM models llm google vision video reasoning agents coding long context

Browse by Model Family