Instructions to use mlx-community/gemma-4-e4b-it-OptiQ-4bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use mlx-community/gemma-4-e4b-it-OptiQ-4bit with MLX:

# Make sure mlx-lm is installed
# pip install --upgrade mlx-lm

# Generate text with mlx-lm
from mlx_lm import load, generate

model, tokenizer = load("mlx-community/gemma-4-e4b-it-OptiQ-4bit")

prompt = "Write a story about Einstein"
messages = [{"role": "user", "content": prompt}]
prompt = tokenizer.apply_chat_template(
    messages, add_generation_prompt=True
)

text = generate(model, tokenizer, prompt=prompt, verbose=True)

Notebooks
Google Colab
Kaggle
Local Apps Settings
LM Studio

How to use mlx-community/gemma-4-e4b-it-OptiQ-4bit with Pi:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "mlx-community/gemma-4-e4b-it-OptiQ-4bit"

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "mlx-lm": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "mlx-community/gemma-4-e4b-it-OptiQ-4bit"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

MLX LM

How to use mlx-community/gemma-4-e4b-it-OptiQ-4bit with MLX LM:

Generate or start a chat session

# Install MLX LM
uv tool install mlx-lm
# Interactive chat REPL
mlx_lm.chat --model "mlx-community/gemma-4-e4b-it-OptiQ-4bit"

Run an OpenAI-compatible server

# Install MLX LM
uv tool install mlx-lm
# Start the server
mlx_lm.server --model "mlx-community/gemma-4-e4b-it-OptiQ-4bit"
# Calling the OpenAI-compatible server with curl
curl -X POST "http://localhost:8000/v1/chat/completions" \
   -H "Content-Type: application/json" \
   --data '{
     "model": "mlx-community/gemma-4-e4b-it-OptiQ-4bit",
     "messages": [
       {"role": "user", "content": "Hello"}
     ]
   }'

Hermes Agent

How to use mlx-community/gemma-4-e4b-it-OptiQ-4bit with Hermes Agent:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "mlx-community/gemma-4-e4b-it-OptiQ-4bit"

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default mlx-community/gemma-4-e4b-it-OptiQ-4bit

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use mlx-community/gemma-4-e4b-it-OptiQ-4bit with OpenClaw:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "mlx-community/gemma-4-e4b-it-OptiQ-4bit"

Configure OpenClaw

# Install OpenClaw:
npm install -g openclaw@latest
# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
  --auth-choice custom-api-key \
  --custom-base-url http://127.0.0.1:8080/v1 \
  --custom-model-id "mlx-community/gemma-4-e4b-it-OptiQ-4bit" \
  --custom-provider-id mlx-lm \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

gemma-4-e4b-it-OptiQ-4bit

Commit History

Oct 18 18 days ago

02:25
Move the vision sidecar to the repo root and add the processor files (LM Studio flattens optiq/; mlx-vlm, oMLX and LM Studio load it from the root)

codelion

Verified signature
b20262a

Sep 14

02:43
Point base_model at the upstream model and mark this repo as a quantization

codelion

Verified signature
e1404a8

Jul 20

01:04
Sync chat template from google/gemma-4-e4b-it (Google canonical, published 2026-07-09)⁠

codelion

Verified signature
6be2eaf

Jul 14

15:46
index: register the bf16 vision tower so stock VLM loaders (mlx-vlm, oMLX, LM Studio) can find it

codelion

Verified signature
4b145b9
09:25
Correct BFCL + Capability: the AST checker could not match array-typed arguments

codelion

Verified signature
2060450
09:08
Correct BFCL + Capability: the AST checker could not match array-typed arguments

codelion

Verified signature
03de9be

Jul 11

02:12
Move OptiQ sidecars under optiq/ so *.safetensors loaders (mlx-vlm, LM Studio) skip them

codelion

Verified signature
a404bb7

Jun 26

02:32
fix: audio_tower conv weights to MLX channel-last layout (full-VLM loader compat)

codelion

Verified signature
6ffaa01

Jun 19

15:32
docs: OptiQ brand spelling

codelion

Verified signature
195e480

Jun 14

02:26
card: remove em-dashes, ensure funnel

codelion

Verified signature
98d7dc6
02:14
card: add mlx-optiq funnel banner + CTA

codelion

Verified signature
63ab053

Jun 6

15:18
Restore vision_config + optiq_vision marker for image input

codelion

Verified signature
fcdb12d
15:18
Add OptIQ vision sidecar (image+text support, v0.2.0)

codelion

Verified signature
7d8118e

May 31

05:56
Add kv_config.json (per-layer mixed-precision KV cache, 5.0 BPW target from OptiQ kv-cache sensitivity analysis). Requires optiq>=0.1.3 runtime for the RotatingQuantizedKVCache shim.

codelion

Verified signature
1ac6d78

May 26

02:40
v0.1.0: 6-metric Capability Score + OptIQ vs U4 deltas

codelion

Verified signature
bef5be4

May 24

10:21
weights re-quant for mlx-lm 0.31 KV-shared Gemma-4 arch: optiq_metadata.json

codelion

Verified signature
8aadccf
10:21
weights re-quant for mlx-lm 0.31 KV-shared Gemma-4 arch: config.json

codelion

Verified signature
5e935ae
10:21
weights re-quant for mlx-lm 0.31 KV-shared Gemma-4 arch: model.safetensors.index.json

codelion

Verified signature
fb6c999
10:21
weights re-quant for mlx-lm 0.31 KV-shared Gemma-4 arch: model-00002-of-00002.safetensors

codelion

Verified signature
2f0bf8a
10:21
weights re-quant for mlx-lm 0.31 KV-shared Gemma-4 arch: model-00001-of-00002.safetensors

codelion

Verified signature
dcecfb9

May 19

13:18
Fix chat template: emit multimodal placeholders in tool messages⁠

prince-canuma

Verified signature
ca4515f

May 10

00:03
card: link blog post for calibration mix instead of bare optiq.jsonl reference

codelion

Verified signature
cfac466

May 9

13:19
card: drop loose IFEval row, keep strict only

codelion

Verified signature
6d66b75

May 7

04:28
v0.1.0 sweep: requant with 5.0-BPW ceiling, fresh benchmark numbers

codelion

Verified signature
13a5b72

Apr 26

05:38
Update Article link to https://x.com/latent_node/status/2028412948167942334

codelion

Verified signature
e60cd68
05:25
Normalize model card: remove twitter ref, lowercase 'optiq', add multimodal stripping note

codelion

Verified signature
7d0b340

Apr 20

15:19
Clean codelion/source refs from model card

codelion

Verified signature
bfe8637
15:11
v0.0.9: fix quant weight round-trip (eager-eval), route to gemma4_text, updated model card

codelion

Verified signature
e49b500
12:11
Strip vision/audio metadata (text-only OptIQ variant cleanup)

codelion

Verified signature
8960b84

Apr 11

16:29
Upload folder using huggingface_hub

codelion

Verified signature
b483edf
16:28
initial commit

codelion

Verified signature
ea0cde4

Commit History

Oct 18 18 days ago

Move the vision sidecar to the repo root and add the processor files (LM Studio flattens optiq/; mlx-vlm, oMLX and LM Studio load it from the root)

Sep 14

Point base_model at the upstream model and mark this repo as a quantization

Jul 20

Sync chat template from google/gemma-4-e4b-it (Google canonical, published 2026-07-09)⁠

Jul 14

index: register the bf16 vision tower so stock VLM loaders (mlx-vlm, oMLX, LM Studio) can find it

Correct BFCL + Capability: the AST checker could not match array-typed arguments

Correct BFCL + Capability: the AST checker could not match array-typed arguments

Jul 11

Move OptiQ sidecars under optiq/ so *.safetensors loaders (mlx-vlm, LM Studio) skip them

Jun 26

fix: audio_tower conv weights to MLX channel-last layout (full-VLM loader compat)

Jun 19

docs: OptiQ brand spelling

Jun 14

card: remove em-dashes, ensure funnel

card: add mlx-optiq funnel banner + CTA

Jun 6

Restore vision_config + optiq_vision marker for image input

Add OptIQ vision sidecar (image+text support, v0.2.0)

May 31

Add kv_config.json (per-layer mixed-precision KV cache, 5.0 BPW target from OptiQ kv-cache sensitivity analysis). Requires optiq>=0.1.3 runtime for the RotatingQuantizedKVCache shim.

May 26

v0.1.0: 6-metric Capability Score + OptIQ vs U4 deltas

May 24

weights re-quant for mlx-lm 0.31 KV-shared Gemma-4 arch: optiq_metadata.json

weights re-quant for mlx-lm 0.31 KV-shared Gemma-4 arch: config.json

weights re-quant for mlx-lm 0.31 KV-shared Gemma-4 arch: model.safetensors.index.json

weights re-quant for mlx-lm 0.31 KV-shared Gemma-4 arch: model-00002-of-00002.safetensors

weights re-quant for mlx-lm 0.31 KV-shared Gemma-4 arch: model-00001-of-00002.safetensors

May 19

Fix chat template: emit multimodal placeholders in tool messages⁠

May 10

card: link blog post for calibration mix instead of bare optiq.jsonl reference

May 9

card: drop loose IFEval row, keep strict only

May 7

v0.1.0 sweep: requant with 5.0-BPW ceiling, fresh benchmark numbers

Apr 26

Update Article link to https://x.com/latent_node/status/2028412948167942334

Normalize model card: remove twitter ref, lowercase 'optiq', add multimodal stripping note

Apr 20

Clean codelion/source refs from model card

v0.0.9: fix quant weight round-trip (eager-eval), route to gemma4_text, updated model card

Strip vision/audio metadata (text-only OptIQ variant cleanup)

Apr 11

Upload folder using huggingface_hub

initial commit