See how ChatGPT, Perplexity and Google AI Overviews describe you today
Free AI Visibility AuditGemma 3 27B is Google's open-weight multimodal model, taking text and image input and returning text, with support for over 140 languages. It is the largest size in the Gemma 3 family and is priced for high-volume work rather than frontier reasoning.
GoogleVerified 24/09/2026Released 12/03/202501 / 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 231 models we track, not against an absolute standard.
Capability
Not measured
No published benchmark scores yet
Input / 1M tokens
$0.08
field median $0.43
Output / 1M tokens
$0.45
field median $1.80
Context
131K
field median 500K
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 27B is a vision-language model built to read text and images and reply in text, with Google citing improved maths, reasoning and chat behaviour over earlier Gemma releases. It is meant for teams who want a small, cheap model they can run at volume across many languages. It is not a frontier reasoning model and should not be treated as a replacement for Google's larger Gemini tier on hard multi-step work.
Where it sits in its line, and when a sibling is the better pick.
Gemma 3 27B first appeared in March 2025 as the top size in the Gemma 3 line, the release that brought multimodality to the family. Pick it when you want the most capable Gemma without moving to a closed frontier model, and when the workload is high-volume classification, extraction, summarising or chat. Pick something larger when the task needs long chains of reasoning or when accuracy matters more than cost per call.
The API, the apps it powers, and what its limits let you do.
Gemma 3 27B is served through standard hosted APIs and, because the weights are open, can also be run on your own hardware. The long context window means you can hand it whole documents, long transcripts or a batch of pages in a single call, and the image input means those pages can be screenshots, scans or charts rather than clean text. Output length is generous enough for full document rewrites and long structured extractions.
What changes because this exists, or why it does not.
Gemma 3 27B makes multimodal work cheap enough to run across an entire corpus rather than a sample. Reading a few hundred thousand scanned pages, screenshots or product images was awkward to justify at frontier pricing, and open weights mean the same pipeline can move on-premise when data cannot leave your estate. The trade is capability at the top end, which is why it sits under Gemini rather than beside it.
Superseded by Gemma 3 12B. 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 | $0.08 | per 1M tokens | 2026-09-24 | Check |
| Output | $0.45 | per 1M tokens | 2026-09-24 | Check |
This is one of the cheapest ways to get image understanding at all, sitting well below the closed frontier models on both input and output and in line with other small open-weight options. At that level, cost stops being the thing you design around and throughput or accuracy becomes the constraint instead.
| Workload | Input / month | Output / month | Cost |
|---|---|---|---|
| A small product team | 20M tokens | 5M tokens | $3.85 |
| A busy support assistant | 200M tokens | 40M tokens | $34.00 |
| A document pipeline | 1000M tokens | 100M tokens | $125.00 |
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 | 117,964 tokens |
| Modalities | text, image |
| Released | 12/03/2025 |
| Status | Current |
| Catalogue identifier | google/gemma-3-27b-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 12B →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
Gemma 3 27B is unlikely to be the model quoting you in an AI answer, so it does not change your citation picture directly. What it does change is the economics of the layer underneath: the tools that read, classify and summarise the web at scale, including the ones your competitors use to monitor how they are described. Assume more of that work is now being done, more often, on more of your content.
Buyers rarely meet Gemma 3 27B by name. It turns up inside products other people built, running the classification, tagging or summarising layer behind a SaaS feature, or self-hosted inside a company that needs its data to stay put. There is no consumer app carrying the Gemma brand, so encounters happen through an API or through somebody else's software.
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.