Instructions to use luisbrel/jimmy-4b-zero-regression with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use luisbrel/jimmy-4b-zero-regression 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("luisbrel/jimmy-4b-zero-regression")

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 luisbrel/jimmy-4b-zero-regression with Pi:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "luisbrel/jimmy-4b-zero-regression"

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": "luisbrel/jimmy-4b-zero-regression"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

MLX LM

How to use luisbrel/jimmy-4b-zero-regression with MLX LM:

Generate or start a chat session

# Install MLX LM
uv tool install mlx-lm
# Interactive chat REPL
mlx_lm.chat --model "luisbrel/jimmy-4b-zero-regression"

Run an OpenAI-compatible server

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

Hermes Agent

How to use luisbrel/jimmy-4b-zero-regression 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 "luisbrel/jimmy-4b-zero-regression"

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 luisbrel/jimmy-4b-zero-regression

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use luisbrel/jimmy-4b-zero-regression with OpenClaw:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "luisbrel/jimmy-4b-zero-regression"

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 "luisbrel/jimmy-4b-zero-regression" \
  --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"

JIMMY-4B-ZERO-REGRESSION (Apple Silicon Native MLX)

luisbrel/jimmy-4b-zero-regression is an experimental neural-network checkpoint developed locally on Apple Silicon M4 Max.

The experiment uses a high-level recursive self-improvement (RSI) loop: candidate versions are repeatedly evaluated against real generation targets, and only promising iterations are retained. It does not use teacher-model distillation or a conventional gradient-descent fine-tuning pipeline.

🏆 Local Head-to-Head Evaluation

Evaluated using authentic autoregressive token generations on Apple Silicon Metal. The run covers 56 generations, eight per suite across two trials:

Benchmark Suite	Base Gemma-4	jimmy-4b-zero-regression	Net Delta ($\Delta$)
GSM8K (Multi-Step Math)	50.0%	75.0%	+25.0%
MATH-500 (Competition Math)	25.0%	37.5%	+12.5%
HumanEval (Python Code)	75.0%	75.0%	0.0%
IFEval (Strict Constraints)	100.0%	100.0%	0.0%
ARC-Challenge (Science)	75.0%	87.5%	+12.5%
Artificial Analysis v5 Multi-Quadrant	100.0%	100.0%	0.0%
GUI & DOM Grounding	62.5%	75.0%	+12.5%
MACRO AVERAGE	69.6%	78.6%	+9.0%

Trial-level scorecards

The aggregate is not enough to judge stability. These are the two trial scores behind each mean:

Benchmark Suite	Base trial 1	Base trial 2	Final trial 1	Final trial 2
GSM8K	75.0%	25.0%	75.0%	75.0%
MATH-500	25.0%	25.0%	50.0%	25.0%
HumanEval	75.0%	75.0%	75.0%	75.0%
IFEval	100.0%	100.0%	100.0%	100.0%
ARC-Challenge	100.0%	50.0%	100.0%	75.0%
Artificial Analysis v5	100.0%	100.0%	100.0%	100.0%
GUI & DOM Grounding	50.0%	75.0%	75.0%	75.0%

💻 Quickstart (MLX Native)

Run on any Mac with Apple Silicon using mlx-lm:

pip install mlx-lm

import mlx_lm

model_path = "luisbrel/jimmy-4b-zero-regression"
model, tokenizer = mlx_lm.load(model_path)

prompt = "If f(x) = (3x-2)/(x-2), calculate f(-2) + f(-1) + f(0). Express your answer as a common fraction."
messages = [{"role": "user", "content": prompt}]
formatted = tokenizer.apply_chat_template(messages, add_generation_prompt=True)

response = mlx_lm.generate(model, tokenizer, prompt=formatted, max_tokens=850)
print(response)

🔬 Provenance & Architecture

Base Architecture: Google Gemma-4-E4B-it (4-bit MLX Quantized)
License: Apache 2.0, inherited from the base model
Runtime: Apple Silicon MLX
Method: Experimental RSI-based post-training
Decontamination: Certified zero 13-gram overlap with benchmark splits

Downloads last month: 22

Safetensors

Model size

7B params

Tensor type

U32

BF16

MLX

Hardware compatibility

4-bit