Open-Source Coding Models
For an open source model for code generation, the strongest matches are meta-llama/codellama (CodeLlama is a specialized family of models built specifically), facebookresearch/codellama (Code Llama is a foundational model specifically trained for) and deepseek-ai/deepseek-coder-v2 (This is a state-of-the-art mixture-of-experts model specifically trained for). qwenlm/qwen2.5 and baichuan-inc/baichuan2 round out the shortlist. Each is ranked by relevance to your query, popularity and recent activity.
我们为您精选了匹配 “best open source coding models” 的开源 GitHub 仓库。结果按与您查询的相关性进行排名 — 您可以使用下方筛选器缩小范围，或通过 AI 进行优化。
Open-Source Coding Models
用 AI 发现最棒的仓库。我们将通过 AI 为您搜索最匹配的仓库。
- meta-llama/codellama
meta-llama/codellama
16,307在 GitHub 上查看CodeLlama is a family of large language models derived from the Llama 2 architecture and specialized for producing, completing, and refactoring source code across multiple programming languages. It functions as a code generation model capable of synthesizing source code from natural language descriptions. The project includes specific model variants designed for different programming tasks. This includes instruction-tuned models trained to follow complex natural language directions and code infilling models that predict and insert missing code segments into existing files by analyzing surroun
CodeLlama is a specialized family of models built specifically for code generation, offering instruction-tuned variants, multi-language support, and large context windows that directly address the requirements for software development tasks.
PythonInstruction Fine-tuningInstruction-Tuned Language ModelsInstruction-Following Models - facebookresearch/codellama
facebookresearch/codellama
16,307在 GitHub 上查看Code Llama is a large language model based on Llama 2 trained specifically for programming tasks and software development. It provides specialized model types optimized for general code generation, instruction following, and context-aware infilling. The project includes an instruction-tuned programming model for executing technical tasks via natural language prompts and a code infilling model that predicts missing sections based on surrounding source context. A large context code model is also provided to analyze extensive blocks of source code for improved coherence. The system covers capab
Code Llama is a foundational model specifically trained for software development tasks, offering instruction-tuned variants, support for large context windows, and specialized capabilities for code infilling and generation.
PythonInstruction Fine-tuningInstruction-Tuned Language ModelsInstruction-Following Models - deepseek-ai/deepseek-coder-v2
deepseek-ai/DeepSeek-Coder-V2
6,462在 GitHub 上查看This is a state-of-the-art mixture-of-experts model specifically trained for code generation and software development, featuring a 128K context window, instruction tuning, and support for fill-in-the-middle objectives.
128K-Token Context WindowsLong Context Processing - qwenlm/qwen2.5
QwenLM/Qwen2.5
27,307在 GitHub 上查看Qwen2.5 is a suite of large language model foundation models designed for natural language generation, code production, and complex mathematical reasoning. The project encompasses a multilingual language model capable of processing dozens of languages and a specialized code generation model for technical problem solving and debugging. The framework is distinguished by its long context capabilities, enabling the analysis of massive inputs ranging from 256K up to 1 million tokens. It further functions as an agentic framework, utilizing standardized templates and parsers to execute autonomous wo
Qwen2.5 is a comprehensive suite of foundation models specifically optimized for code generation and technical reasoning, featuring extensive multi-language support, instruction tuning, and industry-leading context window sizes.
PythonLong Context ProcessingLong-Context ModelsInstruction-Following Models - baichuan-inc/baichuan2
baichuan-inc/Baichuan2
4,098在 GitHub 上查看Baichuan2 is a collection of pre-trained large language models, including base and chat variants, designed for natural language generation and multi-turn conversational AI. It provides an inference engine and a fine-tuning framework to adapt these models to custom datasets and specialized domains. The project features a quantization toolkit and an inference engine that enable model execution across diverse hardware, including graphics processors, central processors, and specialized accelerators. These tools support low-bit weight quantization to reduce memory usage and increase inference spee
This is a general-purpose foundation model that can be fine-tuned for code generation, though it lacks the specialized pre-training for software development found in dedicated coding models.
PythonModel QuantizationModel QuantizationSupervised Instruction Fine-Tuning - facico/chinese-vicuna
Facico/Chinese-Vicuna
4,121在 GitHub 上查看Chinese-Vicuna is a Chinese large language model and instruction-following AI based on the LLaMA architecture. It is specifically designed for natural language understanding and generation in the Chinese language, utilizing an instruction-tuned model to follow complex user prompts across conversations. The project provides a LoRA fine-tuning framework and quantization systems to enable model adaptation and inference on consumer hardware. It implements quantized inference to reduce memory usage on both CPUs and GPUs, supported by a low-level C++ implementation to minimize system resource requi
This is a general-purpose Chinese-language instruction-tuned model that includes support for code-related tasks and low-level inference optimization, though it is not exclusively specialized for software development.
CInstruction Fine-tuningInstruction TuningInstruction-Tuned Language Models - bigcode-project/starcoder
bigcode-project/starcoder
7,508在 GitHub 上查看Starcoder is a large language model and associated framework designed to generate, complete, and evaluate source code across multiple programming languages. It functions as a source code model that can produce complete function implementations and predict subsequent characters in a line of code based on provided prompts. The project provides a specialized toolkit for adapting base models to specific coding tasks and instruction-following behaviors. This includes a conversational code assistant framework for training models to generate code via natural language chat, as well as a parameter-eff
StarCoder is a specialized large language model explicitly designed for code generation and completion across many programming languages, offering the instruction tuning and framework support needed for software development tasks.
PythonInstruction-Tuned Language Models - salesforce/codegen
salesforce/CodeGen
5,175在 GitHub 上查看CodeGen is a trained large language model and program synthesis model designed to generate functional source code. It utilizes a neural network architecture to synthesize executable code from natural language descriptions or partial code snippets. The model enables automated program synthesis and AI-assisted coding by predicting and filling in missing sections of code within a program. It transforms natural language descriptions into functional programming logic to automate the creation of boilerplate and logic.
CodeGen is a specialized large language model designed specifically for program synthesis and code generation, providing the core capability of transforming natural language into functional code as requested.
PythonNatural Language to Code GeneratorsAI Coding AssistantsCausal Language Modeling - salesforce/codet5
salesforce/CodeT5
3,098在 GitHub 上查看Home of CodeT5: Open Code LLMs for Code Understanding and Generation
CodeT5 is a specialized family of open-source models explicitly designed for code understanding and generation tasks, offering the instruction tuning and multi-language support required for software development workflows.
PythonAI Coding AssistantsCode Generation ModelsLarge Language Models - bigcode-project/starcoder2
bigcode-project/starcoder2
2,075在 GitHub 上查看StarCoder2 is a family of code generation models (3B, 7B, and 15B), trained on 600+ programming languages from The Stack v2 and some natural language text such as Wikipedia, Arxiv, and GitHub issues. The models use Grouped Query Attention, a context window of 16,384 tokens, with sliding window…
StarCoder2 is a family of open-source models specifically trained on a massive corpus of code and technical documentation, offering the instruction tuning, multi-language support, and large context window required for high-performance code generation.
PythonLarge Language ModelsLarge Language Models (LLMs)Pre-training Research - ibm-granite/granite-code-models
ibm-granite/granite-code-models
1,250在 GitHub 上查看Granite Code Models is a family of transformer-based foundational models designed for software engineering and logical reasoning tasks. These models are trained on high-quality programming datasets to interpret natural language prompts and generate functional source code, explain complex logic, repair code defects, and produce technical documentation. The project distinguishes itself through specialized training methodologies that align model behavior with complex programming instructions and mathematical problem-solving. By utilizing chain-of-thought reasoning and instruction-tuned parameter
This repository provides a family of foundation models specifically trained for code intelligence tasks, offering instruction-tuned variants that support multiple programming languages and are designed for integration into development workflows.
Instruction TuningInstruction-Tuned Language Models - zai-org/glm-4
zai-org/GLM-4
7,058在 GitHub 上查看GLM-4 is a large language model and fine-tuning framework designed for human-like text production, complex reasoning, and multilingual conversation. It functions as a multimodal system capable of processing high-resolution visual content and as a long-context model designed to analyze documents with a context window of up to one million tokens. The project differentiates itself through a function calling interface that enables AI agent development by connecting the model to external APIs and real-time web browsing. It includes specialized capabilities for generating functional programming cod
GLM-4 is a powerful, instruction-tuned large language model that explicitly supports code generation and long-context processing, making it a strong candidate for software development tasks despite being a general-purpose model rather than one exclusively dedicated to coding.
PythonLong Context ProcessingLong-Context Models - salesforce/codegen2
salesforce/CodeGen2
270在 GitHub 上查看Official research release for the CodeGen2 models (1B, 3B, 7B, 16B) for Program Synthesis as presented in ICLR 2023:
This repository provides a series of pre-trained models specifically designed for program synthesis and code generation, serving as a foundational tool for developers looking to implement or fine-tune code-focused LLMs.
PythonCode Generation Models
