Gemma 2 27B
Google · released Jun 24, 2024 · google/gemma-2-27b-it
- Type
- Open weightsGemma Terms of Use
- Params
- 27.2B
- Context
- 8K
about 6K words of context · download allowed, licence restricts use
Our take
The case for it
- Instruction-following is its strongest measured area: 79.8% on IFEval, which records whether the model does what a prompt asks rather than how well it reasons.
- You can run it yourself rather than depending on a host, so a machine that can hold 27.2 billion parameters can serve it without a hosted offer.
- A long document need not be split before you ask about it, though recall across the whole of it is unverified in our data.
The case against it
- Near the bottom of the preference boards we hold: 151st of 168 on Arena Text (overall) as of 25 Sep 2026, which records which answer people preferred rather than whether it was correct.
- Science and multi-task knowledge are middling: 37.5% correct on GPQA Diamond and 44.5% correct on MMLU-Pro.
- The licence puts conditions on commercial use and redistribution, so it needs reading before you build on it (Gemma Terms of Use).
How good is it?
An open text model from Google for general chat, though it trails most models on everyday questions, writing and code.
- getting answers to everyday questionsArena Text (overall) · 151st of 168
- drafts, rewrites and editingArena Creative Writing · 132nd of 168
- writing and completing codeArena Coding · 157th of 168
EverydayGeneral questions and everyday reasoning
Arena Text (overall)151st of 168 · 1289
CodingWriting and fixing code on its own
Arena Coding157th of 168 · 1304
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 Writing132nd of 168 · 1293
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 model9 scoresEvery figure we hold, from 9 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. 3.7 GB spare means a 10% error in the size would not change the answer.
GeForce RTX 5090 · 32 GB
Room to spare. 11.7 GB spare means a 10% error in the size would not change the answer.
Apple M1 Pro (16-core GPU) · 32 GB
Room to spare. 4.9 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 4 hours ago — each listing carries its own date.
- per 1M tokens
- $0.65 in / $0.65 out
- Context served
- 8K
- Throughput
- Not measured
| Provider | In / out per 1M tokens | Context | Throughput | Trains on prompts | Logs prompts | Zero retention |
|---|---|---|---|---|---|---|
| OpenRouterOpenRouter's own listing | $0.65 / $0.65checked 4 hours ago | 8K | not measured | Unknown | Unknown | Unknown |
| NextBitint4Through OpenRouter | $0.65 / $0.65checked 4 hours ago | 8K2K max reply | 47 tok/s | No | No | Confirmed |
| Google AIDirect | $0.35 / $1.05checked 4 hours ago | not reported8K max reply | not measured | Unknown | Unknown | Unknown |
Across the 3 listings we hold: 1 says it does not train on prompts, 0 say they do and 2 do not say. 1 appears in the zero-retention registry we check; 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 |
|---|---|---|---|
| OpenRouterOpenRouter's own listing | ✗ | ✓ | ✓ |
| NextBitint4Through OpenRouter | ✗ | ✓ | ✓ |
| Google AIDirect |
Tool calling: 0 of 3 listings say yes, 2 say no, 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 2 27B
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.
- 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-2-27b-it
- Takes in, gives back
- Text in, text out
- Catalogue slug
- google-gemma-2-27b