sergiopaniego/qwen3-0.6b-mbpp-grpo-k2

TEXT GENERATIONPricing:Input $0.04 / Cached $0.008 / Output $0.08Concurrent Unit Cost:1Model Size:0.8BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Jul 24, 2026Architecture:Transformer Featherless Exclusive Cold

The sergiopaniego/qwen3-0.6b-mbpp-grpo-k2 model is a 0.8 billion parameter language model, fine-tuned from Qwen/Qwen3-0.6B. It specializes in code generation and problem-solving, having been trained on the MBPP dataset. This model utilizes the GRPO method, enhancing its capabilities for mathematical reasoning and code-related tasks. Its primary strength lies in generating Python code solutions based on natural language prompts.

Loading preview...

Model Overview

This model, sergiopaniego/qwen3-0.6b-mbpp-grpo-k2, is a fine-tuned version of the Qwen3-0.6B architecture, featuring 0.8 billion parameters and a context length of 32768 tokens. It has been specifically optimized for code generation and problem-solving tasks.

Key Capabilities

  • Code Generation: Excels at generating Python code solutions from natural language descriptions.
  • Mathematical Reasoning: Incorporates the GRPO (Gradient-based Reward Policy Optimization) method, as introduced in the DeepSeekMath paper, to enhance its mathematical reasoning abilities, particularly relevant for code-based problems.
  • Fine-tuned on MBPP: Training on the MBPP (Mostly Basic Python Problems) dataset ensures its proficiency in solving programming challenges.

Training Details

This model was trained using the TRL library and the GRPO method. The GRPO method is designed to improve performance in complex reasoning tasks, making this model particularly suitable for scenarios requiring logical and structured outputs, such as code.

Should I use this for my use case?

  • Good for: Generating Python code, solving basic to intermediate programming problems, and tasks requiring structured logical output. Its GRPO training makes it a strong candidate for applications involving mathematical or algorithmic reasoning within a coding context.
  • Consider alternatives if: Your primary use case involves general conversational AI, creative writing, or tasks outside of code generation and structured problem-solving, as it is specialized for programming-related tasks.

Finetunes
1 models