测试开源大型语言模型的游乐场(GPT-OSS, Qwen3.5, DeepSeek)及工具与RAG [免费且无需注册]
📄 中文摘要
提供一个无需注册的在线平台,用户可以在该平台上测试多种流行的开源大型语言模型,包括GPT-OSS、Qwen3.5和DeepSeek。该平台允许用户在自己的硬件上使用vLLM或类似工具运行模型,评估其质量、响应延迟和解码速度,以及RAG摘要能力和工具调用。虽然主要为客户设计以便于决策和测试开源模型,但也向社区开放。用户可以根据需要设置不同的推理努力级别,并可反馈希望添加的模型或功能。
📄 English Summary
Playground to test Open-Source LLMs in action (GPT-OSS, Qwen3.5, DeepSeek) with Tools and RAG [Free and No signup]
An online platform is available for users to test various popular open-source large language models, including GPT-OSS, Qwen3.5, and DeepSeek, without any signup required. Users can run these models on their own hardware using vLLM or similar tools, evaluating their quality, response delay, decoding speed, RAG summarization capabilities, and tool calls. Although primarily designed for clients to facilitate decision-making and testing of open-source models, it is also shared with the community. Users can set different levels of reasoning effort and provide feedback on additional models or features they wish to see.
Powered by Cloudflare Workers + Payload CMS + Claude 3.5
数据源: OpenAI, Google AI, DeepMind, AWS ML Blog, HuggingFace 等