notbad_v1_1_mistral_24b Open Source Large Model - Enhance Mathematical Reasoning and Python Programming Skills

Notbad V1 1 Mistral 24b

Developed by notbadai

A 24B-parameter large language model optimized for mathematical reasoning and Python programming training, based on the Mistral architecture

Large Language Model

Transformers

Open Source License:Apache-2.0 #Mathematical Reasoning Optimization #Python Programming Enhancement #Concise Reasoning Output

Downloads 34

Release Time : 4/7/2025

Model Overview

This model has undergone deep optimization in mathematical reasoning and programming domains, capable of generating more concise reasoning process outputs. It achieves self-iterative optimization through reinforcement learning techniques rather than distillation from other models.

Model Features

Mathematical Reasoning Optimization

Specially reinforced for mathematical problem-solving capabilities

Python Programming Capability

Excellent performance in programming tasks, particularly Python code generation

Concise Reasoning Output

Capable of generating more concise and efficient reasoning processes, reducing redundant information

Self-Iterative Optimization

Achieves capability enhancement through reinforcement learning rather than relying on model distillation

Model Capabilities

Mathematical Problem Solving

Python Code Generation

Text Generation

Logical Reasoning

Instruction Following

Use Cases

Education

Math Tutoring

Helps students understand complex mathematical concepts and problem-solving methods

Scored 0.749 on MMLU Math evaluation

Programming Instruction

Assists in programming learning by providing code examples and explanations

Scored 0.872 on HumanEval evaluation

Development

Code Assistance

Helps developers quickly generate Python code snippets

🚀 Notbad v1.1 Mistral 24B Model Card

This project presents the Notbad v1.1 Mistral 24B model, which significantly outperforms its predecessor, the Notbad v1.0 Mistral 24B, in terms of IFEval scores. It excels in mathematical reasoning and Python coding tasks, offering a more efficient and accurate solution for complex problem - solving.

✨ Features

Enhanced Performance: Achieves better IFEval scores compared to Notbad v1.0 Mistral 24B.
Reasoning Abilities: A reasoning model trained in math and Python coding, with self - improved reasoning capabilities not distilled from other models.
Simplified Output: Capable of producing shorter and cleaner reasoning outputs.
Advanced Training: Built on Mistral - Small - 24B - Instruct - 2501 and further trained with reinforcement learning on math and coding tasks.

📚 Documentation

Model Details

Base Model: Mistral - Small - 24B - Instruct - 2501
Pipeline Tag: text - generation
Library Name: transformers
License: Apache - 2.0

Training Acknowledgments

Special thanks to Lambda and Deep Infra for providing compute resources for research and model training.

Try the Model

You can test the model on chat.labml.ai.

Benchmark Results

General Benchmark

Evaluation	notbad_v1_1_mistral_24b	notbad_v1_0_mistral_24b	mistral - small - 24B - instruct - 2501	gemma - 2b - 27b	llama - 3.3 - 70b	qwen2.5 - 32b	gpt - 4o - mini - 2024 - 07 - 18
mmlu_pro	0.673	0.642	0.663	0.536	0.666	0.683	0.617
gpqa_main	0.467	0.447	0.453	0.344	0.531	0.404	0.377

Math & Coding Benchmark

Evaluation	notbad_v1_0_mistral_24b	notbad_v1_0_mistral_24b	mistral - small - 24B - instruct - 2501	gemma - 2b - 27b	llama - 3.3 - 70b	qwen2.5 - 32b	gpt - 4o - mini - 2024 - 07 - 18
humaneval	0.872	0.869	0.848	0.732	0.854	0.909	0.890
math	0.749	0.752	0.706	0.535	0.743	0.819	0.761

Instruction Following Benchmark

Evaluation	notbad_v1_0_mistral_24b	notbad_v1_0_mistral_24b	mistral - small - 24B - instruct - 2501	gemma - 2b - 27b	llama - 3.3 - 70b	qwen2.5 - 32b	gpt - 4o - mini - 2024 - 07 - 18
ifeval	0.779	0.514	0.829	0.8065	0.8835	0.8401	0.8499

Note: Benchmarks are from Mistral - Small - 24B - Instruct - 2501 Model Card

📄 License

This project is licensed under the Apache - 2.0 license.

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご