44 AI Models.
All on your device.
From lightning-fast tiny models to powerful vision AI. Choose the perfect model for your needs—all running privately on your iPhone.
Filter by Provider
Filter by Use Case
Filter by Size
⭐ Featured Models
Qwen 3 0.6B
Perfect for everyday tasks. Handles casual chat and quick questions.
Qwen 3 VL 2B
Understands images and text. Describe photos, analyze documents, answer visual questions.
Llama 3.2 3B
Balanced model for conversations, writing, and reasoning.
Phi-4 Mini
Microsoft's most advanced small model with outstanding reasoning.
DeepSeek R1 1.5B
Shows reasoning process in thinking tags.
Hermes 3 3B
Exceptional at roleplay and creative writing.
Qwen
8 models
Qwen 3 0.6B
Perfect for everyday tasks. Handles casual chat and quick questions.
Qwen 2.5 0.5B
Lightning fast for basic questions and quick summaries.
Qwen 2.5 1.5B
Perfect everyday assistant that speaks 29+ languages.
Qwen 3 1.7B
Strong reasoning in a compact package.
Qwen 3 VL 2B
Understands images and text. Describe photos, analyze documents, answer visual questions.
Qwen 3 VL 4B
Advanced vision understanding with detailed image analysis.
Qwen 2.5 Coder 1.5B
Compact coding assistant for quick code edits.
Qwen 2.5 Coder 3B
Excellent at code generation and refactoring.
Meta
6 models
Llama 3.2 1B
Fast and efficient for everyday tasks in 8 languages.
Llama 3.2 3B
Balanced model for conversations, writing, and reasoning.
Llama 3.3 1B
Latest generation with improved multilingual support.
Llama 3.3 3B
Enhanced reasoning and instruction following.
Llama 3.2 Vision 11B
Meta's flagship vision model for image understanding.
Llama 3.1 8B
Most capable Llama for complex tasks and long contexts.
Microsoft
3 models
Phi-3 Mini
Exceptional reasoning for its size.
Phi-3.5 Mini
Supports 10+ languages with long-context understanding.
Phi-4 Mini
Microsoft's most advanced small model with outstanding reasoning.
4 models
Gemma 3 270M
Ultra-compact for basic tasks.
Gemma 2 2B
Efficient and capable everyday assistant.
Gemma 2 9B
Google's most capable open model for complex tasks.
PaliGemma 3B
Vision-language model for image understanding.
DeepSeek
4 models
DeepSeek R1 1.5B
Shows reasoning process in thinking tags.
DeepSeek Coder 1.3B
Supports 87 programming languages.
DeepSeek R1 7B
Advanced reasoning rivaling much larger models.
DeepSeek Coder 6.7B
Professional-grade code generation.
Stability AI
3 models
StableLM 2 1.6B
Compact and efficient for edge devices.
StableLM 2 Zephyr 1.6B
Helpful and harmless assistant.
Stable Code 3B
Optimized for code completion.
Mistral AI
1 model
Ministral 3B
Excellent instruction following with multilingual capability.
01.AI
1 model
Yi Coder 1.5B
Supports 52 programming languages.
NVIDIA
1 model
Nemotron Mini 4B
Optimized for chat and function calling.
IBM
3 models
Granite 4.0 350M
Tiny enterprise-focused model.
Granite 4.0 1B
Efficient for business applications.
Granite 4.0 Micro
Handles longer contexts with improved instruction following.
Hugging Face
4 models
SmolLM2 135M
Incredibly tiny for constrained devices.
SmolLM2 360M
Performs above its weight class.
SmolLM2 1.7B
Best SmolLM with remarkable capability.
Zephyr 3B
Helpful assistant fine-tuned from StableLM.
OpenAI
1 model
GPT-2
Text completion model that started it all.
Community
5 models
Hermes 3 3B
Exceptional at roleplay and creative writing.
TinyLlama 1.1B
Remarkably efficient for basic chat.
Dolphin 2.6 3B
Uncensored and creative assistant.
Rocket 3B
Optimized for speed on consumer hardware.
Danube 3 500M
Ultra-efficient for edge devices.
Ready to try these models?
Download KernelAI and start chatting with any of these 44 models privately on your device.
Download KernelAI