{"_id":{"$oid":"6a2301e057a810ff9e91121b"},"id":1,"name":"gpt-oss:20b","description":"An advanced multimodal text generation model powering robust conversational and agentic flows.","parameter":"20","languages":["English"],"supportsFineTuning":true,"modelType":"gpt","modelTypeDescription":"GPT-oss is a multimodal text generation model designed to orchestrate natural language across tasks like writing, coding, summarization, and end-to-end agent workflows.","modelTypeParameterDesc":"agentic text generation tasks","createdAt":"2025-01-10","category":"text generation","provider":"OpenAI"}
{"_id":{"$oid":"6a2301e057a810ff9e91121c"},"id":2,"name":"qwen3:32b","description":"A powerful hybrid reasoning model excelling in agentic coding, vision, and long-context tasks.","parameter":"32","languages":["English","Arabic"],"supportsFineTuning":false,"modelType":"qwen","modelTypeDescription":"Qwen3 is a native vision-language model integrating multi-modal interactive agent capabilities for screen reading and GUI interaction.","modelTypeParameterDesc":"multimodal hybrid reasoning tasks","createdAt":"2026-01-22","category":"text generation","provider":"Alibaba Cloud"}
{"_id":{"$oid":"6a2301e057a810ff9e91121d"},"id":3,"name":"gemma-3:1b","description":"A highly efficient multimodal open model tailored for edge deployment and fast text generation.","parameter":"1","languages":["English"],"supportsFineTuning":true,"modelType":"gemma","modelTypeDescription":"Gemma 3 models handle text and image inputs with a 128K context window, bringing advanced reasoning and function calling to resource-constrained environments.","modelTypeParameterDesc":"edge multimodal tasks","createdAt":"2025-03-12","category":"text generation","provider":"Google DeepMind"}
{"_id":{"$oid":"6a2301e057a810ff9e91121e"},"id":4,"name":"ministral-3:3b","description":"An edge-optimized dense model delivering best-in-class agentic capabilities and vision understanding.","parameter":"3","languages":["English","Arabic"],"supportsFineTuning":true,"modelType":"ministral","modelTypeDescription":"Ministral 3 is a parameter-efficient dense model featuring native function calling, JSON outputting, and a 256K context window for compute-constrained applications.","modelTypeParameterDesc":"agentic edge AI tasks","createdAt":"2026-01-13","category":"text generation","provider":"Mistral AI"}
{"_id":{"$oid":"6a2301e057a810ff9e91121f"},"id":5,"name":"gpt-oss:120b","description":"A massive-scale multimodal model for state-of-the-art enterprise conversational agents and logic reasoning.","parameter":"120","languages":["English","Arabic"],"supportsFineTuning":false,"modelType":"gpt","modelTypeDescription":"GPT-oss offers frontier-level intelligence, supporting highly complex multi-step reasoning, robust agentic workflows, and intricate coding pipelines.","modelTypeParameterDesc":"advanced agentic tasks","createdAt":"2025-06-15","category":"text generation","provider":"OpenAI"}
{"_id":{"$oid":"6a2301e057a810ff9e911220"},"id":6,"name":"qwen3.5:0.8b","description":"A miniature, highly capable model featuring native vision support and hybrid thinking architecture.","parameter":"0.8","languages":["English","Arabic"],"supportsFineTuning":true,"modelType":"qwen","modelTypeDescription":"Qwen3.5 is a multimodal hybrid reasoning LLM family that delivers strong performance for small sizes, excelling in agentic coding and tool-calling.","modelTypeParameterDesc":"lightweight hybrid reasoning tasks","createdAt":"2026-03-05","category":"text generation","provider":"Alibaba Cloud"}
{"_id":{"$oid":"6a2301e057a810ff9e911221"},"id":7,"name":"qwen3.5:2b","description":"A lightweight hybrid reasoning model pushing the Pareto frontier of intelligence for local environments.","parameter":"2","languages":["English","Arabic"],"supportsFineTuning":true,"modelType":"qwen","modelTypeDescription":"Qwen3.5 is a multimodal hybrid reasoning LLM family that delivers strong performance for small sizes, excelling in agentic coding and tool-calling.","modelTypeParameterDesc":"lightweight hybrid reasoning tasks","createdAt":"2026-03-05","category":"text generation","provider":"Alibaba Cloud"}
{"_id":{"$oid":"6a2301e057a810ff9e911222"},"id":8,"name":"qwen3.5:4b","description":"A small-scale vision-language model with highly accurate reasoning and multi-modal interactive capabilities.","parameter":"4","languages":["English","Arabic"],"supportsFineTuning":true,"modelType":"qwen","modelTypeDescription":"Qwen3.5 is a multimodal hybrid reasoning LLM family that delivers strong performance for small sizes, excelling in agentic coding and tool-calling.","modelTypeParameterDesc":"lightweight hybrid reasoning tasks","createdAt":"2026-03-05","category":"text generation","provider":"Alibaba Cloud"}
{"_id":{"$oid":"6a2301e057a810ff9e911223"},"id":9,"name":"qwen3.5:9b","description":"The most intelligent multimodal model in its size class, featuring superior reasoning and agentic tools.","parameter":"9","languages":["English","Arabic"],"supportsFineTuning":true,"modelType":"qwen","modelTypeDescription":"Qwen3.5 is a multimodal hybrid reasoning LLM family that delivers strong performance for small sizes, excelling in agentic coding and tool-calling.","modelTypeParameterDesc":"lightweight hybrid reasoning tasks","createdAt":"2026-03-05","category":"text generation","provider":"Alibaba Cloud"}
{"_id":{"$oid":"6a2301e057a810ff9e911224"},"id":10,"name":"gemma-3:4b","description":"A compact, highly capable multimodal open model optimized for function calling and interactive workflows.","parameter":"4","languages":["English"],"supportsFineTuning":true,"modelType":"gemma","modelTypeDescription":"Gemma 3 models handle text and image inputs with a 128K context window, bringing advanced reasoning and function calling to resource-constrained environments.","modelTypeParameterDesc":"edge multimodal tasks","createdAt":"2025-03-12","category":"text generation","provider":"Google DeepMind"}
{"_id":{"$oid":"6a2301e057a810ff9e911225"},"id":11,"name":"gemma-4:5b","description":"A next-generation portable multimodal model designed for complex agentic orchestration.","parameter":"5","languages":["English"],"supportsFineTuning":true,"modelType":"gemma","modelTypeDescription":"Gemma 4 models expand on function-calling and multi-step logic, offering cutting-edge performance for local device execution.","modelTypeParameterDesc":"advanced edge multimodal tasks","createdAt":"2026-05-01","category":"text generation","provider":"Google DeepMind"}
{"_id":{"$oid":"6a2301e057a810ff9e911226"},"id":12,"name":"gemma-4:8b","description":"A highly advanced portable model excelling at complex visual reasoning and autonomous workflows.","parameter":"8","languages":["English"],"supportsFineTuning":true,"modelType":"gemma","modelTypeDescription":"Gemma 4 models expand on function-calling and multi-step logic, offering cutting-edge performance for local device execution.","modelTypeParameterDesc":"advanced edge multimodal tasks","createdAt":"2026-05-01","category":"text generation","provider":"Google DeepMind"}
{"_id":{"$oid":"6a2301e057a810ff9e911227"},"id":13,"name":"ministral-3:8b","description":"A powerful edge-optimized dense model delivering state-of-the-art vision and native JSON function calling.","parameter":"8","languages":["English","Arabic"],"supportsFineTuning":true,"modelType":"ministral","modelTypeDescription":"Ministral 3 is a parameter-efficient dense model featuring native function calling, JSON outputting, and a 256K context window for compute-constrained applications.","modelTypeParameterDesc":"agentic edge AI tasks","createdAt":"2026-01-13","category":"text generation","provider":"Mistral AI"}