{"generated_at":"2026-10-23T04:00:03+08:00","query":{"keyword":"","gguf":false,"source":"","limit":50},"total":263,"count":50,"pull_hint":"zlm pull <pull>","models":[{"name":"embeding","pull":"zctech/x-ai-agi:embeding","desc":"尊创文本向量模型（307M）","caps":["embedding"],"sizes":["307m"],"source":"zctech","license":"ZCTECH"},{"name":"img2text","pull":"zctech/x-ai-agi:img2text","desc":"尊创图文识别模型（1.1B）","caps":["vision"],"sizes":["1.1b"],"source":"zctech","license":"ZCTECH"},{"name":"zai3turbo","pull":"zctech/x-ai-agi:zai3turbo","desc":"尊创 9.4B 通用推理模型（Q8_0，长上下文）","caps":["tools","completion"],"sizes":["9.4b"],"source":"zctech","license":"ZCTECH"},{"name":"zai4turbo","pull":"zctech/x-ai-agi:zai4turbo","desc":"尊创自研 9.0B 通用推理模型（Q6_K，支持工具调用与思考）","caps":["tools","thinking"],"sizes":["9.0b"],"source":"zctech","license":"ZCTECH","gguf":true},{"name":"zai5turbo","pull":"zctech/x-ai-agi:zai5turbo","desc":"尊创 27.3B 大模型（双卡 24G×2 推荐）","caps":["tools","thinking"],"sizes":["27.3b"],"source":"zctech","license":"ZCTECH"},{"name":"alfred","pull":"alfred","desc":"A robust conversational model designed to be used for both chat and instruct use cases.","caps":[],"sizes":["40b"],"source":"public","gguf":true,"pulls":"307.9K","tags":7,"updated":"2 years ago"},{"name":"all-minilm","pull":"all-minilm","desc":"Embedding models on very large sentence level datasets.","caps":["embedding"],"sizes":["22m","33m"],"source":"public","gguf":true,"pulls":"3.8M","tags":10,"updated":"2 years ago"},{"name":"athene-v2","pull":"athene-v2","desc":"Athene-V2 is a 72B parameter model which excels at code completion, mathematics, and log extraction tasks.","caps":["tools"],"sizes":["72b"],"source":"public","gguf":true,"pulls":"607.2K","tags":17,"updated":"1 year ago"},{"name":"aya","pull":"aya","desc":"Aya 23, released by Cohere, is a new family of state-of-the-art, multilingual models that support 23 languages.","caps":[],"sizes":["8b","35b"],"source":"public","gguf":true,"pulls":"1.1M","tags":33,"updated":"2 years ago"},{"name":"aya-expanse","pull":"aya-expanse","desc":"Cohere For AI&#39;s language models trained to perform well across 23 different languages.","caps":["tools"],"sizes":["8b","32b"],"source":"public","gguf":true,"pulls":"1.1M","tags":33,"updated":"1 year ago"},{"name":"bakllava","pull":"bakllava","desc":"BakLLaVA is a multimodal model consisting of the Mistral 7B base model augmented with the LLaVA architecture.","caps":["vision"],"sizes":["7b"],"source":"public","gguf":true,"pulls":"907.4K","tags":17,"updated":"2 years ago"},{"name":"beam","pull":"beam","desc":"Beam is an open model by Reflection designed to deliver strong performance for areas such as agentic tool use, coding and general reasoning with an emphasis on token efficient usage.","caps":["tools","thinking"],"sizes":[],"source":"public","gguf":true,"pulls":"29.6K","tags":null,"updated":"6 days ago"},{"name":"bespoke-minicheck","pull":"bespoke-minicheck","desc":"A state-of-the-art fact-checking model developed by Bespoke Labs.","caps":[],"sizes":["7b"],"source":"public","gguf":true,"pulls":"547K","tags":17,"updated":"2 years ago"},{"name":"bge-large","pull":"bge-large","desc":"Embedding model from BAAI mapping texts to vectors.","caps":["embedding"],"sizes":["335m"],"source":"public","gguf":true,"pulls":"295K","tags":3,"updated":"2 years ago"},{"name":"bge-m3","pull":"bge-m3","desc":"BGE-M3 is a new model from BAAI distinguished for its versatility in Multi-Functionality, Multi-Linguality, and Multi-Granularity.","caps":["embedding"],"sizes":["567m"],"source":"public","gguf":true,"pulls":"7.9M","tags":3,"updated":"2 years ago"},{"name":"bonsai-2","pull":"bonsai-2","desc":"Bonsai 2 from Prism ML is a family of vision models built for ternary weights – smaller models with close to full-precision quality.","caps":["vision","tools","thinking"],"sizes":["2b","4b","9b","27b"],"source":"public","gguf":true,"pulls":"43K","tags":9,"updated":"6 days ago"},{"name":"clef","pull":"clef","desc":"Clef is a 27B multimodal model created by Cloudflare that turns a state and a schema of typed questions into decisions.","caps":["vision","decision"],"sizes":["27b"],"source":"public","gguf":true,"pulls":"33K","tags":9,"updated":"1 week ago"},{"name":"clef-flash","pull":"clef-flash","desc":"Clef-Flash is a 9B multimodal model that turns a state and a schema of typed questions into decisions.","caps":["vision","decision"],"sizes":["9b"],"source":"public","gguf":true,"pulls":"35.2K","tags":5,"updated":"1 week ago"},{"name":"codebooga","pull":"codebooga","desc":"A high-performing code instruct model created by merging two existing code models.","caps":[],"sizes":["34b"],"source":"public","gguf":true,"pulls":"517.6K","tags":16,"updated":"2 years ago"},{"name":"codegeex4","pull":"codegeex4","desc":"A versatile model for AI software development scenarios, including code completion.","caps":[],"sizes":["9b"],"source":"public","gguf":true,"pulls":"708.5K","tags":17,"updated":"2 years ago"},{"name":"codegemma","pull":"codegemma","desc":"CodeGemma is a collection of powerful, lightweight models that can perform a variety of coding tasks like fill-in-the-middle code completion, code generation, natural language understanding, mathematical reasoning, and instruction following.","caps":[],"sizes":["2b","7b"],"source":"public","gguf":true,"pulls":"3.3M","tags":85,"updated":"2 years ago"},{"name":"codellama","pull":"codellama","desc":"A large language model that can use text prompts to generate and discuss code.","caps":[],"sizes":["7b","13b","34b","70b"],"source":"public","gguf":true,"pulls":"7.3M","tags":199,"updated":"2 years ago"},{"name":"codeqwen","pull":"codeqwen","desc":"CodeQwen1.5 is a large language model pretrained on a large amount of code data.","caps":[],"sizes":["7b"],"source":"public","gguf":true,"pulls":"1.1M","tags":30,"updated":"2 years ago"},{"name":"codestral","pull":"codestral","desc":"Codestral is Mistral AI’s first-ever code model designed for code generation tasks.","caps":[],"sizes":["22b"],"source":"public","gguf":true,"pulls":"1.4M","tags":17,"updated":"2 years ago"},{"name":"codeup","pull":"codeup","desc":"Great code generation model based on Llama2.","caps":[],"sizes":["13b"],"source":"public","gguf":true,"pulls":"613K","tags":19,"updated":"2 years ago"},{"name":"cogito","pull":"cogito","desc":"Cogito v1 Preview is a family of hybrid reasoning models by Deep Cogito that outperform the best available open models of the same size, including counterparts from LLaMA, DeepSeek, and Qwen across most standard benchmarks.","caps":["tools"],"sizes":["3b","8b","14b","32b","70b"],"source":"public","gguf":true,"pulls":"2.1M","tags":20,"updated":"1 year ago"},{"name":"cogito-2.1","pull":"cogito-2.1","desc":"The Cogito v2.1 LLMs are instruction tuned generative models. All models are released under MIT license for commercial use.","caps":[],"sizes":["671b"],"source":"public","gguf":true,"pulls":"230.3K","tags":5,"updated":"11 months ago"},{"name":"command-a","pull":"command-a","desc":"111 billion parameter model optimized for demanding enterprises that require fast, secure, and high-quality AI","caps":["tools"],"sizes":["111b"],"source":"public","gguf":true,"pulls":"237.5K","tags":5,"updated":"1 year ago"},{"name":"command-r","pull":"command-r","desc":"Command R is a Large Language Model optimized for conversational interaction and long context tasks.","caps":["tools"],"sizes":["35b"],"source":"public","gguf":true,"pulls":"1.6M","tags":32,"updated":"2 years ago"},{"name":"command-r-plus","pull":"command-r-plus","desc":"Command R&#43; is a powerful, scalable large language model purpose-built to excel at real-world enterprise use cases.","caps":["tools"],"sizes":["104b"],"source":"public","gguf":true,"pulls":"825.3K","tags":21,"updated":"2 years ago"},{"name":"command-r7b","pull":"command-r7b","desc":"The smallest model in Cohere&#39;s R series delivers top-tier speed, efficiency, and quality to build powerful AI applications on commodity GPUs and edge devices.","caps":["tools"],"sizes":["7b"],"source":"public","gguf":true,"pulls":"340.7K","tags":5,"updated":"1 year ago"},{"name":"command-r7b-arabic","pull":"command-r7b-arabic","desc":"A new state-of-the-art version of the lightweight Command R7B model that excels in advanced Arabic language capabilities for enterprises in the Middle East and Northern Africa.","caps":["tools"],"sizes":["7b"],"source":"public","gguf":true,"pulls":"211.1K","tags":5,"updated":"1 year ago"},{"name":"d1","pull":"d1","desc":"d1 is a 3B decision model from Liquid AI, built on LFM2.5-VL-3B","caps":["vision"],"sizes":["3b"],"source":"public","gguf":true,"pulls":"18.5K","tags":3,"updated":"1 week ago"},{"name":"d1-omni","pull":"d1-omni","desc":"d1-omni is a multimodal 600M decision model from Liquid AI","caps":["vision"],"sizes":["600m"],"source":"public","gguf":true,"pulls":null,"tags":4,"updated":"4 days ago"},{"name":"dbrx","pull":"dbrx","desc":"DBRX is an open, general-purpose LLM created by Databricks.","caps":[],"sizes":["132b"],"source":"public","gguf":true,"pulls":"376.6K","tags":7,"updated":"2 years ago"},{"name":"deepcoder","pull":"deepcoder","desc":"DeepCoder is a fully open-Source 14B coder model at O3-mini level, with a 1.5B version also available.","caps":[],"sizes":["1.5b","14b"],"source":"public","gguf":true,"pulls":"967.1K","tags":9,"updated":"1 year ago"},{"name":"deepscaler","pull":"deepscaler","desc":"A fine-tuned version of Deepseek-R1-Distilled-Qwen-1.5B that surpasses the performance of OpenAI’s o1-preview with just 1.5B parameters on popular math evaluations.","caps":[],"sizes":["1.5b"],"source":"public","gguf":true,"pulls":"1.3M","tags":5,"updated":"1 year ago"},{"name":"deepseek-coder","pull":"deepseek-coder","desc":"DeepSeek Coder is a capable coding model trained on two trillion code and natural language tokens.","caps":[],"sizes":["1.3b","6.7b","33b"],"source":"public","gguf":true,"pulls":"4.7M","tags":102,"updated":"2 years ago"},{"name":"deepseek-coder-v2","pull":"deepseek-coder-v2","desc":"An open-source Mixture-of-Experts code language model that achieves performance comparable to GPT4-Turbo in code-specific tasks.","caps":[],"sizes":["16b","236b"],"source":"public","gguf":true,"pulls":"3.3M","tags":64,"updated":"2 years ago"},{"name":"deepseek-llm","pull":"deepseek-llm","desc":"An advanced language model crafted with 2 trillion bilingual tokens.","caps":[],"sizes":["7b","67b"],"source":"public","gguf":true,"pulls":"1.3M","tags":64,"updated":"2 years ago"},{"name":"deepseek-ocr","pull":"deepseek-ocr","desc":"DeepSeek-OCR is a vision-language model that can perform token-efficient OCR.","caps":["vision"],"sizes":["3b"],"source":"public","gguf":true,"pulls":"561.5K","tags":3,"updated":"11 months ago"},{"name":"deepseek-r1","pull":"deepseek-r1","desc":"DeepSeek-R1 is a family of open reasoning models with performance approaching that of leading models, such as O3 and Gemini 2.5 Pro.","caps":["tools","thinking"],"sizes":["1.5b","7b","8b","14b","32b","70b","671b"],"source":"public","gguf":true,"pulls":"94M","tags":35,"updated":"1 year ago"},{"name":"deepseek-v2","pull":"deepseek-v2","desc":"A strong, economical, and efficient Mixture-of-Experts language model.","caps":[],"sizes":["16b","236b"],"source":"public","gguf":true,"pulls":"1.2M","tags":34,"updated":"2 years ago"},{"name":"deepseek-v2.5","pull":"deepseek-v2.5","desc":"An upgraded version of DeekSeek-V2 that integrates the general and coding abilities of both DeepSeek-V2-Chat and DeepSeek-Coder-V2-Instruct.","caps":[],"sizes":["236b"],"source":"public","gguf":true,"pulls":"299.8K","tags":7,"updated":"2 years ago"},{"name":"deepseek-v3","pull":"deepseek-v3","desc":"A strong Mixture-of-Experts (MoE) language model with 671B total parameters with 37B activated for each token.","caps":[],"sizes":["671b"],"source":"public","gguf":true,"pulls":"3.8M","tags":5,"updated":"1 year ago"},{"name":"deepseek-v3.1","pull":"deepseek-v3.1","desc":"DeepSeek-V3.1-Terminus is a hybrid model that supports both thinking mode and non-thinking mode.","caps":["tools","thinking"],"sizes":["671b"],"source":"public","gguf":true,"pulls":"740.1K","tags":7,"updated":"1 year ago"},{"name":"deepseek-v4-pro","pull":"deepseek-v4-pro","desc":"DeepSeek-V4-Pro is a frontier Mixture-of-Experts model with a large context window and three reasoning modes.","caps":["tools","thinking"],"sizes":[],"source":"public","gguf":true,"pulls":"466.7K","tags":2,"updated":"2 months ago"},{"name":"deepseek-v4.1-flash","pull":"deepseek-v4.1-flash","desc":"Designed for greater capability, faster inference, higher throughput, and scaling to larger models.","caps":["vision","tools","thinking"],"sizes":[],"source":"public","gguf":true,"pulls":"165.2K","tags":2,"updated":"5 days ago"},{"name":"devstral","pull":"devstral","desc":"Devstral: the best open source model for coding agents","caps":["tools"],"sizes":["24b"],"source":"public","gguf":true,"pulls":"1.1M","tags":5,"updated":"1 year ago"},{"name":"devstral-2","pull":"devstral-2","desc":"123B model that excels at using tools to explore codebases, editing multiple files and power software engineering agents.","caps":["tools"],"sizes":["123b"],"source":"public","gguf":true,"pulls":"367.5K","tags":5,"updated":"10 months ago"}]}