Gemma 3 12B
Google · released Mar 1, 2025 · google/gemma-3-12b-it
- Type
- Open weightsGemma Terms of Use
- Params
- 12.2B
- Context
- 131K
about 98K words of context · download allowed, licence restricts use
Our take
Written Sep 30, 2026Gemma 3 12B is a mid-size model you can download and run on a single modern graphics card, or reach through a host for very little. It handles text and images and takes long documents in one request, but the preference boards place it in the bottom quarter of the field, so treat it as a workhorse rather than a leader.
Use it for everyday chat, summarising and questions about screenshots where the bill matters more than peak quality, and where you want the option of running the model yourself. Read the Gemma Terms of Use before you build a commercial product on it. Skip it if a task needs measured coding or reasoning evidence, or if answer quality is the deciding factor.
The case for it
- Cheap to run through a host: the cheapest listed offer sits well under the next tier up, so a high-volume workload costs little.
- Pictures go in with the question, so a screenshot or a diagram does not have to be described in words first.
- The request capacity takes a long report or a stack of documents beside the question, though reliable recall across all of it is unverified in our data.
The case against it
- 130th of 168 on Arena Text (overall) as of 25 Sep 2026 and 151st of 168 on Arena Coding as of 25 Sep 2026, so it is not the pick when answer quality decides.
- The Gemma Terms of Use put conditions on commercial use and redistribution, so a commercial product needs the terms checked first.
- Every score we hold is a human preference rating, which records which answer people liked rather than whether it was correct, so coding and reasoning ability need a trial on work you can check yourself.
How good is it?
An open text model for general chat and writing, though it trails most models on everyday questions and coding.
- getting answers to everyday questionsArena Text (overall) · 130th of 168
- writing and completing codeArena Coding · 151st of 168
EverydayGeneral questions and everyday reasoning
Arena Text (overall)130th of 168 · 1342
CodingWriting and fixing code on its own
Arena Coding151st of 168 · 1316
Arena Coding is the only board that has scored it for this.
AgenticPlanning, calling tools, staying on task
Not yet scored on Arena Agent.
WritingDrafting and rewriting prose
Arena Creative Writing111th of 168 · 1332
Arena Creative Writing is the only board that has scored it for this.
These tests check whether a model follows instructions — a precondition for all the work above, but not a measure of how well that work is done.
Every published score for this model6 scoresEvery figure we hold, from 6 boards, with who ran it and a link to the source — including the boards no rating above is built on.
Can you run it yourself?
Comfortable fit
GeForce RTX 4090 · 24 GB
Room to spare. 13.6 GB spare means a 10% error in the size would not change the answer.
GeForce RTX 5090 · 32 GB
Room to spare. 21.6 GB spare means a 10% error in the size would not change the answer.
Apple M1 (8-core GPU) · 16 GB
Room to spare. 2.8 GB spare means a 10% error in the size would not change the answer.
Memory use by level
Against a 24 GB card.
What is quantisation? →This model on every device we track71 devicesThe Q4 build most people download, on each device: what the weights come to, how much context the memory leaves, and whether it runs. Smallest device that runs it first.
Check against your own machine → · Where to rent it hosted →
Or rent it from someone else
Prices checked 2 hours ago — each listing carries its own date.
- per 1M tokens
- $0.050 in / $0.10 out
- Context served
- 131K
- Throughput
- Not measured
| Provider | In / out per 1M tokens | Context | Throughput | Trains on prompts | Logs prompts | Zero retention |
|---|---|---|---|---|---|---|
| Novita AIDirect | $0.050 / $0.10checked 2 hours ago | 131K | not measured | Unknown | Unknown | Unknown |
| OpenRouterOpenRouter's own listing | $0.050 / $0.15checked 2 hours ago | 131K | not measured | Unknown | Unknown | Unknown |
| DeepInfrabf16Direct and through OpenRouter | $0.050 / $0.15checked 2 hours ago | 131K16K max reply through OpenRouter | 30 tok/sthrough OpenRouter | DirectUnknownThrough OpenRouterNo | DirectUnknownThrough OpenRouterNo | DirectUnknownThrough OpenRouterConfirmed |
OpenRouter
DeepInfraAcross the 3 listings we hold: 1 says it does not train on prompts (only through OpenRouter), 0 say they do and 2 do not say. 1 appears in the zero-retention registry we check (only through OpenRouter); the rest are unknown to us.
What each host's API supports
From the parameter list each endpoint publishes. Streaming is omitted: nothing we hold reports it, for any model.
| Provider | Tool calling | JSON output | Strict schema |
|---|---|---|---|
| Novita AIDirect | |||
| OpenRouterOpenRouter's own listing | ✓ | ✓ | ✓ |
| DeepInfrabf16Direct and through OpenRouter | ✓ | ✓ | ✓ |
Tool calling: 2 of 3 listings say yes, 1 publishes no parameter list. JSON output: 2 of 3 listings say yes, 1 publishes no parameter list. Strict schema: 2 of 3 listings say yes, 1 publishes no parameter list.
Models people weigh against Gemma 3 12B
When we formed this view
Recent changes
What moved
first indexed by our pipelineEach date is the day we first saw the change, or the day the maker announced it.
What we do not know about this model yet
- We hold no measured file for it, so all 3 sizes on this page are calculated from the parameter count.
- 1 of 3 listings publishes no parameter list, so what its API accepts is unknown to us.
- Nothing we hold says whether an endpoint streams, so we do not show it either way.
- 2 of 3 listings do not say whether they train on prompts, and 1 answers only through OpenRouter, not for its own listing.
- We hold no cached-input rate for any of its listings.
- We hold no batch or off-peak rate for any of its listings.
- We hold a decode speed for it, but no prompt-processing (prefill) figure, so how long the input side of a job takes is unknown to us.
Licence and identifiers
What the licence allowsGemma Terms of Use, what it allows commercially, and the identifiers you need to pull this model — its Hugging Face repo, our slug and a machine-readable card.
Licence
Gemma Terms of Use
Commercial use allowed, but Google's prohibited-use policy applies and can be updated over time — terms are less static than Apache/MIT.
Identifiers
- Hugging Face
- google/gemma-3-12b-it
- Takes in, gives back
- Text and images in, text out
- Catalogue slug
- google-gemma-3-12b