News, models and guides from Modelbase.https://modelbase.app/https://modelbase.app/blog/ask-ai-about-a-photo-on-iphone/https://modelbase.app/blog/ask-ai-about-a-photo-on-iphone/Snap a plant, a menu or a sign and ask what it is. How to ask an on-device AI about a picture in Modelbase, which models can see images, and what to ask.Fri, 13 Nov 2026 00:00:00 GMThttps://modelbase.app/blog/best-local-llm-apps-for-mac/https://modelbase.app/blog/best-local-llm-apps-for-mac/LM Studio, Ollama, Jan, Msty, AnythingLLM, GPT4All and App Store apps like Modelbase and Private LLM: how the main ways to run AI models on a Mac compare, as of November 2026.Fri, 13 Nov 2026 00:00:00 GMThttps://modelbase.app/blog/chat-with-pdf-on-iphone/https://modelbase.app/blog/chat-with-pdf-on-iphone/Ask questions about a PDF, summarize it or find a clause with an AI that reads the file on your iPhone, iPad or Mac. The steps in Modelbase, good prompts, and the limits.Fri, 13 Nov 2026 00:00:00 GMThttps://modelbase.app/blog/best-offline-ai-apps-for-iphone/https://modelbase.app/blog/best-offline-ai-apps-for-iphone/Modelbase, PocketPal AI, Private LLM, Locally AI, Liquid Apollo and Enclave run AI models on your iPhone. Price, models, photos, voice and privacy compared, as of November 2026.Fri, 13 Nov 2026 00:00:00 GMThttps://modelbase.app/blog/chatgpt-offline/https://modelbase.app/blog/chatgpt-offline/No: ChatGPT needs a connection because it runs on OpenAI's servers. Here's how to get a private AI chat that works offline on your iPhone, iPad or Mac.Fri, 13 Nov 2026 00:00:00 GMThttps://modelbase.app/blog/free-llm-api/https://modelbase.app/blog/free-llm-api/Which LLM APIs still have a free tier: OpenRouter's free models, Groq and Google's Gemini API, and what changed at Mistral, Cerebras, Hugging Face and GitHub. Limits, data use, and how to use them on iPhone and Mac.Fri, 13 Nov 2026 00:00:00 GMThttps://modelbase.app/blog/gemini-api-free-tier/https://modelbase.app/blog/gemini-api-free-tier/The Gemini API has a free tier with Gemini Flash models, Gemini 4 Hydrogen and Gemma 4, as of November 2026. What's free, how to get a key in Google AI Studio, what Google does with free-tier data, and how to connect it on iPhone and Mac.Fri, 13 Nov 2026 00:00:00 GMThttps://modelbase.app/blog/gemma-vs-llama-vs-qwen/https://modelbase.app/blog/gemma-vs-llama-vs-qwen/Google's Gemma, Meta's Llama and Qwen compared as on-device models: sizes, photos, thinking, context and which iPhone each suits, with every model in the Modelbase catalog.Fri, 13 Nov 2026 00:00:00 GMThttps://modelbase.app/blog/groq-api-free-tier/https://modelbase.app/blog/groq-api-free-tier/Groq's API has a free plan: gpt-oss-120b, gpt-oss-20b and Qwen 3.8 27B, 1,000 requests a day each as of November 2026. The limits, what Groq does with your data, and how to use it on iPhone, iPad and Mac.Fri, 13 Nov 2026 00:00:00 GMThttps://modelbase.app/blog/lm-studio-vs-ollama/https://modelbase.app/blog/lm-studio-vs-ollama/LM Studio and Ollama both run open AI models on your own computer for free. How they differ in interface, license, models, MLX support and API, as of November 2026.Fri, 13 Nov 2026 00:00:00 GMThttps://modelbase.app/blog/llama-3-2-on-iphone-and-mac/https://modelbase.app/blog/llama-3-2-on-iphone-and-mac/Meta's Llama 3.2 1B and 3B are small text models built for phones. Their sizes, context length, languages and license, what they're good at, and how to run them offline on iPhone, iPad and Mac.Fri, 13 Nov 2026 00:00:00 GMThttps://modelbase.app/blog/lfm-2-5-liquid-ai/https://modelbase.app/blog/lfm-2-5-liquid-ai/LFM 2.5 is Liquid AI's family of small open models built for running on devices. What the 350M, 1.2B, Thinking and vision models are, what they're good at, and how to run them on iPhone, iPad and Mac.Fri, 13 Nov 2026 00:00:00 GMThttps://modelbase.app/blog/is-chatgpt-safe/https://modelbase.app/blog/is-chatgpt-safe/ChatGPT is a legitimate, widely used service, but it isn't private by default. Does ChatGPT store your data, train on it or share it? What OpenAI says, the settings to change, and how to keep chats on your own device.Fri, 13 Nov 2026 00:00:00 GMThttps://modelbase.app/blog/ollama-gui-apps/https://modelbase.app/blog/ollama-gui-apps/Ollama runs open models on your computer; these apps give it a chat window. Ollama's own app, Open WebUI, Msty, AnythingLLM and iPhone clients, compared as of November 2026.Fri, 13 Nov 2026 00:00:00 GMThttps://modelbase.app/blog/qwen-vs-deepseek/https://modelbase.app/blog/qwen-vs-deepseek/On a phone, the DeepSeek you can run is built on Qwen. How Qwen 3.5, Qwen 3 and DeepSeek R1 (1.5B) compare in size, thinking, vision and context, and which to pick for what.Fri, 13 Nov 2026 00:00:00 GMThttps://modelbase.app/blog/openrouter-free-models/https://modelbase.app/blog/openrouter-free-models/Is OpenRouter free? Its :free models cost nothing, with daily limits. Which free models there are in November 2026, the limits, what happens to your data, and how to use them on iPhone, iPad and Mac.Fri, 13 Nov 2026 00:00:00 GMThttps://modelbase.app/blog/on-device-ai-vs-cloud-ai/https://modelbase.app/blog/on-device-ai-vs-cloud-ai/On-device AI runs on your phone or computer; cloud AI runs in a data center. How they differ in privacy, offline use, cost and capability, and when each makes sense.Fri, 13 Nov 2026 00:00:00 GMThttps://modelbase.app/blog/private-chatgpt-alternatives/https://modelbase.app/blog/private-chatgpt-alternatives/Want ChatGPT's features without sending your chats to a server? Apps that run AI models on your own device, matched to each ChatGPT feature, as of November 2026.Fri, 13 Nov 2026 00:00:00 GMThttps://modelbase.app/blog/phi-4-mini/https://modelbase.app/blog/phi-4-mini/Phi-4-mini is Microsoft's 3.8B open model, strong at math, logic and following instructions. Its size, context, languages and limits, how it differs from Phi-4, and how to run it offline on iPhone, iPad and Mac.Fri, 13 Nov 2026 00:00:00 GMThttps://modelbase.app/blog/run-deepseek-r1-locally-on-iphone-and-mac/https://modelbase.app/blog/run-deepseek-r1-locally-on-iphone-and-mac/Run DeepSeek R1 offline on your iPhone or Mac: what the 1.5B distilled model can and can't do, the steps in Modelbase, and other reasoning models to try.Fri, 13 Nov 2026 00:00:00 GMThttps://modelbase.app/blog/run-llm-locally-on-iphone-and-mac/https://modelbase.app/blog/run-llm-locally-on-iphone-and-mac/Run open models like Qwen, Gemma, Llama and DeepSeek on your own iPhone, iPad or Mac: what you need, how to pick a model that fits, and the steps.Fri, 13 Nov 2026 00:00:00 GMThttps://modelbase.app/blog/small-language-models/https://modelbase.app/blog/small-language-models/Small language models (SLMs) are AI models small enough to run on a phone or laptop. What counts as small, what 1B and 4-bit mean, and real examples with sizes.Fri, 13 Nov 2026 00:00:00 GMThttps://modelbase.app/blog/voice-chat-with-ai/https://modelbase.app/blog/voice-chat-with-ai/Have a spoken conversation with an AI that runs on your iPhone: on-device speech recognition, a local model and natural voices. How voice mode works in Modelbase, and how to set it up.Fri, 13 Nov 2026 00:00:00 GMThttps://modelbase.app/blog/what-is-a-reasoning-model/https://modelbase.app/blog/what-is-a-reasoning-model/A reasoning model writes out its thinking before it answers. How reasoning models work, when thinking helps, and what Instant, Medium and High mean on iPhone and Mac.Fri, 13 Nov 2026 00:00:00 GMThttps://modelbase.app/blog/what-is-a-token-in-ai/https://modelbase.app/blog/what-is-a-token-in-ai/AI models read and write text in tokens, small pieces of words. What a token is, how many words fit in one, what a context window is, and why a phone's is smaller than the model's.Fri, 13 Nov 2026 00:00:00 GMThttps://modelbase.app/blog/what-is-mlx/https://modelbase.app/blog/what-is-mlx/MLX is Apple's open-source machine learning framework for Apple silicon. What it does, why unified memory matters, MLX vs GGUF, and how apps use it.Fri, 13 Nov 2026 00:00:00 GMThttps://modelbase.app/blog/which-model-on-iphone/https://modelbase.app/blog/which-model-on-iphone/How to pick a local LLM for your iPhone, iPad or Mac: what fits in memory, and the best small models from Qwen, Gemma, Llama, Phi and DeepSeek.Fri, 13 Nov 2026 00:00:00 GMThttps://modelbase.app/blog/introducing-modelbase/https://modelbase.app/blog/introducing-modelbase/Modelbase is a free app that runs open AI models on your iPhone, iPad and Mac with Apple's MLX: offline, private, no account, first model included.Thu, 12 Nov 2026 00:00:00 GMT