# Code Generation

DiffuCoder-7B-cpGRPO-6bit is a text generation model converted based on the MLX format, focusing on code and text diffusion tasks.

The GGUF quantized version of Qwen2.5-Coder-1.5B, optimized for code generation tasks, offering multiple quantization options to balance performance and resource consumption.

Llamacpp imatrix quantized version based on WhiteRabbitNeo-V3-7B, specializing in cybersecurity and DevOps tasks with code generation support.

Seed-Coder-8B-Reasoning is an open-source code model with a scale of 8B, focusing on code generation and reasoning tasks, with powerful performance and efficient parameter utilization.

A text generation and inference model optimized using Unsloth and TRL libraries based on the Qwen/Qwen2.5-Coder-7B model, achieving 2x faster training speed

Seed-Coder-8B-Reasoning is a large language model with 8B parameters developed by ByteDance-Seed, focusing on code generation and reasoning tasks.

Llamacpp quantized version based on andrewzh's Absolute_Zero_Reasoner-Coder-7b model, supporting multiple quantization levels, suitable for reasoning and code generation tasks.

Qwen2.5-Coder-7B-Instruct is a 7B-parameter code generation model based on the Qwen2.5 architecture, specializing in instruction fine-tuning, suitable for code generation and programming assistance tasks.

An instruction-tuned model based on the Qwen3-14B architecture, specializing in mathematical reasoning, code generation, and factual accuracy

Falcon-H1 is an efficient hybrid architecture language model developed by TII, combining the strengths of Transformers and Mamba architectures, supporting English and multilingual tasks.

Seed-Coder-8B-Reasoning is an 8B-scale open-source code model with enhanced reasoning capabilities through reinforcement learning, supporting a context length of 65,536 and excelling in programming tasks.

TableLLM is a large language model specifically designed for handling table data manipulation tasks, catering to real-world office scenarios involving table data processing.

Gemma 3 is a lightweight open model series built by Google based on Gemini technology, supporting multimodal input and text output, featuring a 128K large context window and support for 140+ languages.

Phi-4-reasoning-plus is an advanced open-weight reasoning model developed by Microsoft Research, optimized through supervised fine-tuning and reinforcement learning based on Phi-4, focusing on advanced reasoning capabilities in mathematics, science, and coding fields.

DeepCoder-14B-Preview is a code generation model developed based on DeepSeek-R1-Distill-Qwen-14B, focusing on solving verifiable programming problems.

Phi-4 Reasoning is a cutting-edge open-weight reasoning model based on Phi-4, fine-tuned with supervised chain-of-thought trajectory datasets and trained via reinforcement learning, specializing in mathematics, science, and programming skills.

DeepSeek-V3-0324-Pruned-Coder-411B is a pruned and optimized code generation model based on the DeepSeek-V3 architecture, focusing on code generation tasks.

The fifth-generation Qwen2.5-YOYO model integrates features from multiple advanced models, optimizes the model merging formula, and supports a context length of 1 million tokens.

A large language model fine-tuned on the codeforces-cots dataset based on google/gemma-3-12b-it

OlympicCoder-7B is a 7B-parameter large language model focused on code generation, based on open-r1/OlympicCoder-7B with llama.cpp quantization, supporting multiple quantization level options.

Quantized version of OlympicCoder-32B, based on llama.cpp's imatrix quantization method, suitable for code generation tasks.

Kanana is a series of bilingual large language models developed by Kakao, excelling in Korean tasks while maintaining competitiveness in English tasks. The 2.1b version is the lightweight base model of this series.

YuLan-Mini-Instruct is a compact yet powerful 2.4-billion-parameter text generation model, specializing in mathematical and code reasoning tasks with support for both English and Chinese.

This is a GGUF format model converted from the Qwen2.5-Coder-0.5B model, suitable for code generation tasks.

Yi-Coder is a series of open-source code language models that achieve state-of-the-art coding performance with fewer than 10 billion parameters.

FP8 quantized version of DeepSeek-Coder-V2-Lite-Instruct, suitable for commercial and research use in English, optimized for inference efficiency.

An orthogonalized version of Codestral-22B-v0.1, where the model's rejection capability has been removed through ablation techniques to better align with user requests.

A code generation model fine-tuned based on distilgpt2, suitable for code-related tasks

Phi-3-Mini-4K-Instruct is a 3.8 billion parameter lightweight cutting-edge open-source model trained on the Phi-3 dataset, emphasizing high quality and inference-intensive characteristics.

Phi-3 Mini 128K Instruct is a 3.8B parameter lightweight open-source model focused on reasoning capabilities, supporting 128K context length.

Arctic is a dense Mixture of Experts (MoE) architecture large language model developed by the Snowflake AI Research team, with 480 billion parameters, open-sourced under the Apache-2.0 license.

A Python programming assistant model fine-tuned based on google/gemma-2b, specializing in Python code generation and problem-solving

CodeGemma is a lightweight open-source collection of code models based on Gemma, specializing in code generation, completion, and conversational tasks.

Code Llama is a series of code generation and understanding models released by Meta, ranging from 7 billion to 34 billion parameters. This version is the 34 billion parameter base model.

Code Llama is a series of pre-trained generative text models released by Meta, focusing on code generation and understanding, with versions ranging from 7 billion to 34 billion parameters.

Code Llama is a series of pretrained and fine-tuned generative text models developed by Meta, with parameter scales ranging from 7B to 34B, suitable for general code generation and understanding

Code Llama is a series of code generation and comprehension models released by Meta, including pre-trained and fine-tuned versions with parameters ranging from 7B to 34B. This model is the 7B-parameter instruction fine-tuned version, specifically optimized for code assistant scenarios.

Code Llama is a series of code generation and understanding models from Meta with parameter scales ranging from 7B to 34B. This version is the 7B base model.

A multi-domain language model fine-tuned on the Hyperion-v2.0 dataset, excelling in scientific reasoning and complex task processing.

A large language model fine-tuned based on the starcoder2-3b model, specializing in code generation tasks, achieving a score of 65.9 pass@1 in the HumanEval-Python test