---

license: apache-2.0

license_link: https://ai.google.dev/gemma/docs/gemma_4_license

base_model: google/gemma-4-E2B-it-qat-q4_0-unquantized

base_model_relation: quantized

library_name: gguf

pipeline_tag: text-generation

tags:

• gemma
• gguf
• llama.cpp

---

Gemma 4 E2B Artemis GGUF

Start here. Download gemma-4-E2B-it-Artemis-Q3_K_L.gguf (2.31 GB). 95.8% same next tokens as BF16; Unsloth UD-Q4_K_XL is closer if a 2.6 GB file fits. Text only. Details.

<!-- chart:start -->

!Gemma 4 E2B: Same next token as BF16 (%) versus file size. Each point shows a measured file. Exact values are in the comparison tables below.

Next-token agreement with BF16 on 640 scored positions, measured on an AMD Ryzen 9 9950X CPU with llama.cpp 0504396. BF16 agrees with itself by definition. Token agreement is not task accuracy; smaller variants drift more in longer continuations. No bartowski result is reported. The unconfirmed-checkpoint row is omitted from the plot; all results remain in the comparison table.

<!-- chart:end -->

Earlier file names still work for now; new downloads should use the names below. Use the exact filename or the :Q3_K_L tag; without a tag, apps may pick a different file.

llama-server -hf turintech/gemma-4-E2B-it-Artemis-GGUF:Q3_K_L --port 8080 -c 8192 --jinja

<div>

<a href="https://turintech.ai"><img src="https://huggingface.co/turintech/gemma-4-E2B-it-Artemis-GGUF/resolve/main/assets/turintech-logo.png" width="180" alt="TurinTech"></a>

<p style="margin: 4px 0;"><em>TurinTech models, built with Artemis</em></p>

</div>

Gemma 4 E2B is Google's smallest Gemma 4 instruction model for chat, reasoning and tool use, built

for laptops and phones. These files are made from Google's quantisation-aware (QAT) checkpoint

(google/gemma-4-E2B-it-qat-q4_0-unquantized),

which Google trained to hold up well at low precision. Text only.

This is an independent TurinTech conversion, not affiliated with or endorsed by Google. The files are standard GGUF

and run on stock llama.cpp. Changes from the original: the weights are converted to GGUF and quantized for

llama.cpp. See NOTICE.

Why use this file

• Run locally so inputs can stay on the same machine.
• Work offline after downloading the model and runtime.
• Use for local chat, summaries and code assistance.
• Choose a compact GGUF download; file size is not a measurement of runtime memory.

Files

| Variant | Size | Saved | Next token | First 32 | Acc. Δ | Choose for |

| --- | --- | --- | --- | --- | --- | --- |

| BF16 original (QAT) | 9.3 GB | | 100.0% | reference | reference | |

| Q3_K_L | 2.31 GB | 75% | 95.8% | 74.7% | -1.6 | Default. |

| Q3_K_M | 2.20 GB | 76% | 93.1% | 69.0% | +0.3 | Smaller download. |

| IQ3_XXS | 2.05 GB | 78% | 91.6% | 67.1% | 0.0 | Smallest. |

"Saved" is the file-size reduction versus the BF16 original. "Acc. Δ" is the accuracy change in percentage points; the other columns are defined below. Variant links download the exact GGUF; full filenames are in the commands and checksums. File labels are size classes, explained under "File names, sizes and SHA-256 checksums" at the end of the card.

Q3_K_L: Default. Closest to the original among our files. Loses 3 of 16 maths questions in our sample.

"Same next token" is how often a file predicts the same next token as the original. "Matching tokens (first 32)" is the

share of the first 32 generated tokens that match the original's token at the same position (greedy decoding,

averaged over 50 prompts). "Task

accuracy change" is the difference in accuracy against the original on 310 task items, in points. This small sample does not establish equivalence. KL (in the test table) measures how far a file's next-token probabilities are from

the original's; lower is closer.

Run it

llama-server -m gemma-4-E2B-it-Artemis-Q3_K_L.gguf --port 8080 -c 8192 --jinja

Tested on an Apple M3 Max (36 GB) with llama.cpp Metal (b11617): loads and answers with the command above; about 115 tokens/s generation with the Q3_K_L file.

curl http://localhost:8080/v1/chat/completions -H 'Content-Type: application/json' -d '{
  "messages": [{"role": "user", "content": "In two sentences, why is the sky blue?"}],
  "max_tokens": 1500
}'

-c 8192 sets an 8,192-token context; --jinja uses the chat template stored in the file. The template turns on

thinking by default, so allow enough output tokens (about 1,500) or turn thinking off in the request.

# a smaller file: download it, then the same command
hf download turintech/gemma-4-E2B-it-Artemis-GGUF gemma-4-E2B-it-Artemis-Q3_K_M.gguf --local-dir .
llama-server -m gemma-4-E2B-it-Artemis-Q3_K_M.gguf --port 8080 -c 8192 --jinja

Questions, problems or a model to consider for a smaller version? Reply in our feedback thread.

Run it with Ollama

The same files are packaged on Ollama with the Gemma 4 renderer and parser:

ollama run turintech/gemma-4-e2b-it-artemis:2.3gb

Ollama tags are named by size: 2.3gb (the default, also latest) is the Q3_K_L file, 2.2gb is Q3_K_M and 2.1gb is IQ3_XXS. Each tag holds the same file, with the same SHA-256, as this repository. To import the GGUF locally instead, use Ollama's local import with the exact default file:

hf download turintech/gemma-4-E2B-it-Artemis-GGUF gemma-4-E2B-it-Artemis-Q3_K_L.gguf --local-dir .
printf 'FROM ./gemma-4-E2B-it-Artemis-Q3_K_L.gguf\nPARAMETER num_ctx 8192\n' > Modelfile
ollama create gemma-4-e2b-it-artemis -f Modelfile
ollama run gemma-4-e2b-it-artemis

This recipe follows Ollama's GGUF import instructions. It sets an 8,192-token context. The quality results on this card were measured with llama.cpp; importing the file into Ollama does not establish equivalent results or compatibility with every Ollama version.

<details>

<summary>Evaluation details and limitations</summary>

How it was tested

Recommendation notes

Download gemma-4-E2B-it-Artemis-Q3_K_L.gguf (2.31 GB) and run it with llama.cpp's chat server, as in

Run it. It predicts the same next token as the original 95.8% of the time and loses 3 of 16 maths questions

in our task sample. If a 2.6 GB file fits, Unsloth's UD-Q4_K_XL is closer to the original. Memory use has not been

measured yet. Text only.

These files are conversions of Google's quantization-aware Gemma checkpoint. Google also publishes its own Q4_0 GGUF of it (google/gemma-4-E2B-it-qat-q4_0-gguf), which is not in the comparisons below. Compare the measured results below with the named public alternatives when choosing a file.

All three files keep Google's Q4_0 tensors and store the per-layer embedding table at lower precision (Q3_K, IQ3_XXS or IQ2_S); their labels are size classes, explained in the file names table at the end.

Use the exact filename or the :Q3_K_L tag in the commands above; without a tag, an automatic download may select a different file.

Size and quality trade-offs

Google's Gemma 4 E2B is 9.3 GB at full precision (BF16). The files here are

2.1 to 2.3 GB, 75 to 78% smaller. The Artemis Q3_K_L file predicts the same next token as the original 95.8% of the time. On a 310-item task sample none of the

three shows a statistically detectable change (the Artemis Q3_K_L file: -1.6 points, 95% range -5.6 to +2.4; it loses 3 of 16

maths questions). Below 2.6 GB, the public files we tested that come from the same checkpoint predict a different next

token 45 to 55% of the time and lose 23 to 34 points on the same tasks.

If a 2.6 GB file fits the available storage, Unsloth's UD-Q4_K_XL is closer to the original (table below). These files are for

when a task needs a smaller download.

Every file, ours and public, was compared with the BF16 conversion of the same QAT original at identical llama.cpp

settings (context 8192, flash attention on, F16 KV cache, greedy decoding, thinking off), on:

• KL divergence and next-token agreement over 640 positions (20 fixed documents, 32 scored positions each),
• greedy 32-token continuations on 50 prompts (10 each of code repair, maths, chat, multilingual and report excerpts),
• a 310-item task sample from ARC-Challenge, WinoGrande, HellaSwag, MMLU-Pro, GSM8K, BFCL and IFEval (64 ARC-Challenge, 64 WinoGrande, 64 HellaSwag, 70 MMLU-Pro, 16 GSM8K, 16 BFCL, 16 IFEval).

This is a sample for comparing files, not a benchmark score.

| File | Size | KL | Same next token | Matching tokens (first 32) | Task accuracy change |

|---|---|---|---|---|---|

| mradermacher i1-IQ3_XXS | 2.65 GB | .594 | 64.4% | 25.4% | -9.4 |

| Unsloth UD-Q4_K_XL (QAT) | 2.62 GB | .0014 | 98.3% | 89.1% | -0.3 |

| mradermacher i1-IQ2_M | 2.55 GB | 1.84 | 49.1% | 7.5% | -26.5 |

| mradermacher i1-IQ2_S | 2.50 GB | 1.20 | 54.5% | 14.9% | -23.5 |

| mradermacher i1-IQ2_XXS | 2.42 GB | 1.94 | 45.0% | 5.2% | -34.2 |

| Artemis Q3_K_L | 2.31 GB | .016 | 95.8% | 74.7% | -1.6 |

| Artemis Q3_K_M | 2.20 GB | .022 | 93.1% | 69.0% | +0.3 |

| Unsloth UD-Q2_K_XL (source checkpoint not confirmed, so not like for like) | 2.19 GB | 1.35 | 52.0% | 11.4% | -14.8 |

| Artemis IQ3_XXS | 2.05 GB | .037 | 91.6% | 67.1% | 0.0 |

Public files: unsloth/gemma-4-E2B-it-qat-GGUF revision 66a399f, mradermacher/gemma-4-E2B-it-qat-q4_0-unquantized-i1-GGUF

revision 62734e9. Measured with llama.cpp 0504396 on an AMD Ryzen 9 9950X, CPU only (8 threads). Every row comes from

the same machine. No speed or memory numbers yet.

</details>

Memory

Peak memory has not been measured for these files. Download size is not a measurement of the memory needed to run them; context length and runtime settings also affect memory use.

<details>

<summary>File names, sizes and SHA-256 checksums</summary>

File names

File names end in a standard llama.cpp quant label, so llama.cpp, Ollama and LM Studio can pick a file by tag. All three files keep Google's Q4_0 tensors and store the per-layer embedding table at lower precision, so their labels are size classes: the smallest standard llama.cpp quant whose typical bits per weight are at least the file's. The general.file_type field inside every file reads Q4_0.

| File | Earlier name | Bits per weight | Main tensor types (share of bytes) |

|---|---|---|---|

| gemma-4-E2B-it-Artemis-Q3_K_L.gguf | gemma-4-E2B-it-Artemis-2.3GB.gguf | 3.96 | Q4_0 56%, Q3_K 44% |

| gemma-4-E2B-it-Artemis-Q3_K_M.gguf | gemma-4-E2B-it-Artemis-2.2GB.gguf | 3.77 | Q4_0 59%, IQ3_XXS 41% |

| gemma-4-E2B-it-Artemis-IQ3_XXS.gguf | gemma-4-E2B-it-Artemis-2.1GB.gguf | 3.52 | Q4_0 63%, IQ2_S 37% |

The earlier names still download the same bytes, with the same checksums, for a transition period.

File checksums

| File | Bytes | sha256 |

|---|---|---|

| gemma-4-E2B-it-Artemis-Q3_K_L.gguf | 2,308,419,584 | 1e207a449e37ea902de237aba1158f960a91bd344bee6edc5b0aa0dbce015621 |

| gemma-4-E2B-it-Artemis-Q3_K_M.gguf | 2,198,319,104 | 6bdb901ce3c468c626c5481076b9c56bc69a35fc3ad08c78b2b6cbede5cbd9da |

| gemma-4-E2B-it-Artemis-IQ3_XXS.gguf | 2,051,518,464 | 66b7eebe3316775b6e0194b3b4da2888c32830d3c7e3c749b1e861d121e97aa9 |

</details>

Custom versions

We also make fine-tuned and further optimised versions on request: tuned on specific data, built for the target hardware

(CPU servers, laptops, Apple Silicon, edge devices), or compressed from a private model. Book a call with TurinTech.

Licence

Apache-2.0, the licence Google releases Gemma 4 under (Gemma 4 licence page); see LICENSE and

NOTICE.

When serving these files to others, for example through an API, an inference provider or OpenRouter, we'd

appreciate a mention of TurinTech (turintech.ai) as the source of the quantization.

These files keep improving.