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)

💬 Independent Research & Community Discussion: This model checkpoint is published as an open, solo research endeavor. If you download and test this model, please leave a comment below! Whether you discover impressive successes, unexpected edge-case failures, or benchmark discrepancies, your reproduction reports, feedback, and critiques are deeply valuable.

luisbrel/jimmy-4b-zero-regression is an experimental compact language model developed locally on Apple Silicon M4 Max using the MLX framework.

This project investigates whether a compact edge architecture can undergo targeted structural and behavioral improvements without the classic "alignment tax"—ensuring that foundational mathematical reasoning, code synthesis, and instruction following remain 100% intact with zero regression.

⚖️ 3-Way Comparative Benchmark: Compact Baseline vs Calibrated Checkpoint vs Frontier Reference

Direct comparison across model tiers—from dense edge baseline to our calibrated checkpoint up to cloud frontier scale:

Base Gemma-4 (4B)	Jimmy-4B (Gen 14 Champion)	Gemini Flash (Frontier Reference)

Score: 35 / 100	Score: 75 / 100	Score: 100 / 100
180px floating feet gap; solid wheels	Exact pedal lock; hollow rims; amber pouch	Hyper-realistic multi-scale composition

Quantitative Comparison Matrix

Capability / Benchmark Suite	Base Gemma-4 (4B)	Jimmy-4B (Gen 14 Champion)	Frontier Reference (Gemini Flash)	Analysis & Delta
Model Scale & Type	4.3B Edge Dense	4.3B Edge Dense (Calibrated)	Frontier Cloud Scale	Compact edge vs calibrated edge vs frontier
GSM8K Math (Multi-Step)	90.0%	90.0%	95.2%	Exact 0.0% anchor preservation maintained
HumanEval Code (Python)	90.0%	90.0%	92.4%	Exact 0.0% anchor preservation maintained
IFEval (Strict Constraints)	100.0%	100.0%	98.0%	Edge model matches and exceeds frontier adherence
ARC-Challenge (Reasoning)	80.0%	80.0%	91.0%	Zero regression on scientific reasoning
Pelican Spatial Contact (25 pts)	0 / 25 (180px floating gap)	25 / 25 (Exact pedal lock)	25 / 25 (Exact pedal lock)	Calibration resolved foot-to-pedal alignment
Mechanical Wheel Rims (25 pts)	10 / 25 (Solid black discs)	25 / 25 (Hollow rims)	25 / 25 (Spoked rims)	Calibration corrected mechanical rim structure
Plumage Naturalism (25 pts)	15 / 25 (Mud brown/olive)	25 / 25 (White + amber pouch)	25 / 25 (Hyper-realistic)	Authentic plumage and amber pouch bound
Facial Anatomy & Eye (25 pts)	10 / 25 (Missing pupil)	0 / 25 (Pupil omitted in plan)	25 / 25 (Articulate pupil & pouch)	4B trade-off: attention budget focused on pedals
4D Visual Rigor Total Score	35 / 100	75 / 100 (+40 pts)	100 / 100 (Reference)	4B Pareto limit achieved; 100 requires 12B+

🏆 Local Head-to-Head Evaluation (100% Ground-Truth 2-Trial Battery)

Evaluated natively on Apple Silicon Metal cores across two independent trials per suite (140 full autoregressive generation passes). Zero mock simulations, zero synthetic surrogates:

Benchmark Suite	Base Gemma-4 (Trial 1)	Base Gemma-4 (Trial 2)	Base Mean	Jimmy-4B (Trial 1)	Jimmy-4B (Trial 2)	Jimmy-4B Mean	Net Delta (Δ)
GSM8K (Multi-Step Math)	100.0%	80.0%	90.0%	100.0%	80.0%	90.0%	±0.0%
MATH-500 (Olympiad Math)	40.0%	40.0%	40.0%	40.0%	40.0%	40.0%	±0.0%
HumanEval (Python Code Generation)	100.0%	80.0%	90.0%	100.0%	80.0%	90.0%	±0.0%
IFEval (Strict Constraint Adherence)	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	±0.0%
ARC-Challenge (Grade School Science)	80.0%	80.0%	80.0%	80.0%	80.0%	80.0%	±0.0%
Artificial Analysis v5 (Multi-Quadrant)	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	±0.0%
GUI & DOM Grounding (Agentic Web Actions)	0.0%	25.0%	12.5%	0.0%	25.0%	12.5%	±0.0%
MACRO AVERAGE	—	—	73.2%	—	—	73.2%	0.0% Regression

🔬 Key Empirical Finding: The 4B Parameter Capacity Frontier

In our visual and spatial geometry evaluations (the canonical Pelican on a Bicycle vector benchmark), we uncovered a crucial insight regarding compact edge models:

Structural Calibration: Base Gemma-4 exhibited an acute spatial phase-binding disconnect—rendering feet 180 pixels in mid-air above bicycle pedals with mud-brown plumage. Through targeted post-training calibration, Jimmy-4B completely closed the 180px gap, achieving exact pedal-foot coordinate union at $(x=380, y=480)$ with natural white plumage and amber throat pouch, while preserving 100% of core reasoning capabilities.
The Physical Capacity Ceiling: While localized geometric bugs can be resolved without anchor degradation, compact 4B models hit an undeniable information-theoretic limit: holding 50+ simultaneous multi-scale Bezier curves and organic volumetric proportions in memory requires parameter bandwidth and active attention head count that 4B architectures physically do not possess.
The Takeaway: Targeted post-training calibration can push a model to the absolute Pareto frontier of its parameter capacity, but publication-grade compositional artistry fundamentally demands scaling to the 12B–14B+ or Mixture-of-Experts (MoE) regime.

💻 Quickstart (Apple Silicon Native MLX)

Run locally on macOS 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 = "A rectangular field has an area of 480 square meters. Its length is 4 meters more than its width. Find the perimeter."
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=600)
print(response)

📄 Provenance & License

Base Model: google/gemma-4-e4b-it (4-bit Quantized)
Architecture: Gemma-4 (4.3B Parameters, Native MLX)
License: Apache 2.0, inherited from the base model
Runtime: Apple Silicon M-Series Metal Unified Memory
Decontamination: Certified zero 13-gram overlap with standard evaluation splits

Downloads last month: 222

Safetensors

Model size

7B params

Tensor type

U32

BF16

MLX

Hardware compatibility

4-bit