See how ChatGPT, Perplexity and Google AI Overviews describe you today
Free AI Visibility AuditGemma 3 12B is Google's mid-sized open model, built to take text and images in and return text out, with support for over 140 languages.
GoogleVerified 24/09/2026Released 13/03/202501 / decision
Each figure sits next to the middle of the field, so it can be read as dear or cheap, wide or narrow, rather than floating on its own. Compared against the 179 models we track, not against an absolute standard.
Capability
Not measured
No published benchmark scores yet
Input / 1M tokens
$0.05
field median $0.46
Output / 1M tokens
$0.15
field median $1.82
Context
131K
field median 524K
02 / overview
Four questions, answered separately, because somebody arrives at one of them rather than at the top of the page.
The job it was built for, and the job it is not for.
Gemma 3 12B is an open-weight vision-language model: it reads text and images and writes text back. Google positions it for general chat, reasoning and maths work across a wide spread of languages. It does not generate images or audio, and it is not the frontier tier of Google's line-up, so the heaviest research and long-form agentic work belongs elsewhere.
Where it sits in its line, and when a sibling is the better pick.
Gemma 3 12B first appeared in March 2025 as part of the Gemma 3 release, the generation that brought multimodal input to the family. The 12B sits in the middle of the range, the sensible pick when the smaller Gemma sizes drop too much quality but the larger one costs more to serve than the job warrants. If the task is straightforward classification or extraction at volume, step down; if it needs deep multi-step reasoning, step up.
The API, the apps it powers, and what its limits let you do.
Gemma 3 12B is available through the usual hosted inference providers and, because the weights are open, can be run on your own infrastructure. The long context is enough to hold a full documentation set, a long support thread or a batch of screenshots in a single call, and the image input means you can send product UI, charts or scanned pages alongside the prompt rather than transcribing them first.
What changes because this exists, or why it does not.
An open model that handles images and long documents at this price changes what is worth automating. Work that was too expensive to run over an entire content library, tagging, translating, summarising, checking pages against a brief, becomes routine batch work, and the open weights mean it can run inside your own environment when the data cannot leave.
Superseded by Gemma 3 4B. See the whole line.
03 / evidence
Split by category, so a strong number never hides a thin evidence base. Verified means we read it on the benchmark's own published results; a provider's claim about its own model is shown and labelled rather than dropped.
No published benchmark scores for this model yet.
We publish a score only where we can link the result to where it was published. Until a benchmark result for this model exists in a source we read, this section stays empty rather than being filled with a provider's marketing figure.
How we decide what counts as evidence04 / ledger
Every published row, grouped by category, each compared with the best published score on the same benchmark. 'Is 64% good' is a question nobody can answer; '26 points behind the leader' is one anybody can.
Nothing in the ledger yet.
Each row here carries a score, the benchmark it came from, what the leading model scored on the same test, and a link to the published result. Rows appear as results are published and read.
How we decide what counts as evidence05 / capability
Where this model is strong, and against how many peers. Ranks are against models with evidence in that category, not against everything we track: ranking against models nobody tested would rank who published, not who is better.
No category scores to shape yet.
A category score is the weighted mean of the benchmarks published for it. With no published rows there is nothing to average, and an empty chart drawn at zero would say something false.
How we decide what counts as evidence06 / cost
List API rates as last read from the provider, with the source on every row, plus every change we have recorded since we started tracking it.
| Charge | Price | Unit | Read on | Source |
|---|---|---|---|---|
| Input | $0.05 | per 1M tokens | 2026-09-24 | Check |
| Output | $0.15 | per 1M tokens | 2026-09-24 | Check |
Gemma 3 12B is among the cheaper options a team can call through an API, low enough that running it across every page, ticket or image in an archive is a reasonable line item rather than a project to justify. Because the weights are open, the hosted price is also a ceiling: self-hosting trades the per-token cost for infrastructure you already run.
| Workload | Input / month | Output / month | Cost |
|---|---|---|---|
| A small product team | 20M tokens | 5M tokens | $1.75 |
| A busy support assistant | 200M tokens | 40M tokens | $16.00 |
| A document pipeline | 1000M tokens | 100M tokens | $65.00 |
List API rates, no caching and no batch discount, which both providers offer and which change the answer a great deal. Treat these as the ceiling, not the bill.
07 / specs
As listed by the provider's own catalogue and re-read every few hours. Anything absent is absent there too.
| Context window | 131,072 tokens |
|---|---|
| Maximum output | 16,384 tokens |
| Modalities | text, image |
| Released | 13/03/2025 |
| Status | Current |
| Catalogue identifier | google/gemma-3-12b-it |
08 / lineage
What this model replaced, what replaced it, and what else its provider has in the field.
Also from Google
09 / line
Every model in this family in release order, so a page from eight months ago says in one glance that two newer ones exist.
Came before
The first we track in this line.
Came after
Gemma 3 4B →Ordered by release date and worked out from the naming, so a new member slots in as soon as its page exists. A retired model keeps its page and its place in the line.
10 / notes
What this model changes for a brand trying to be cited in AI answers, and every change we have logged since it launched.
What it changes for you
Cheap open models push the cost of large-scale content processing down far enough that competitors can audit, rewrite and expand their whole site continuously. If your category is being summarised by systems running on models like this, the winning position goes to the brand whose pages are clearest to parse, not the one with the most pages. Structure, consistent naming and plain factual claims are what survive that kind of bulk reading.
Buyers rarely meet Gemma 3 12B by name. They meet it inside products built on it: support assistants, in-app summarisers, document tools and internal search built by teams who wanted multimodal input without a frontier bill. It is an API and self-host model rather than a consumer chat app.
Nothing published here yet. Changes appear within hours of a provider announcing them.
11 / questions
The things people ask about this model, answered from what is on this page rather than from anywhere else.
Models change what gets cited. We measure whether AI answers name your brand or your competitors across every assistant, and show you what to change.