MODELS BY RAM

Best local AI models for 16 GB RAM

With 16 GB of RAM, you can run most models up to 3–4B parameters at good quality. This tier covers reasoning models like Phi-4 Mini and a wide range of chat, code, and vision models.

54 models run · 1 too heavy · High tier
FILTER BY TASK
Chat 19 models
Bonsai 8B (1-bit)
Bonsai 8B (1-bit) 1.16 GB
PrismML's native 1-bit model — 8B params in 1.16 GB, needs forked llama.cpp
Q1 ~86 tok/s
Tight fit → Qwen 2.5 0.5B
Qwen 2.5 0.5B 477 MB
Lightweight chat — surprisingly capable
FP16 ~83 tok/s
Runs great → Llama 3.2 1B
Llama 3.2 1B 791 MB
Meta's solid all-rounder
FP16 ~37 tok/s
Runs great → Gemma 3 1B
Gemma 3 1B 786 MB
Google's compact model — great reasoning for its size
FP16 ~45 tok/s
Runs great → 💬
LFM2.5 1.2B 731 MB
Liquid AI's hybrid model — blazing fast CPU inference
FP16 ~43 tok/s
Runs great → SmolLM2 360M
SmolLM2 360M 267 MB
Ultra-small language model
FP16 ~109 tok/s
Runs great → SmolLM2 135M
SmolLM2 135M 99 MB
Ultra-tiny language model — runs anywhere
FP16 ~200 tok/s
Runs great → Llama 3.1 8B
Llama 3.1 8B 4.9 GB
Meta's workhorse 8B — excellent all-around
Q8 ~12 tok/s
Runs well → Qwen 2.5 7B
Qwen 2.5 7B 4.4 GB
Top-tier 7B chat — rivals much larger models
Q8 ~12 tok/s
Runs well → Llama 3.2 3B
Llama 3.2 3B 2.0 GB
Meta's 3B — best small model for many tasks
FP16 ~15 tok/s
Runs great → TinyLlama 1.1B
TinyLlama 1.1B 780 MB
Popular tiny model trained on 3T tokens
FP16 ~45 tok/s
Runs great → SmolLM2 1.7B
SmolLM2 1.7B 1.0 GB
Largest SmolLM2 — punches above its weight
FP16 ~29 tok/s
Runs great → Qwen 2.5 3B
Qwen 2.5 3B 2.0 GB
Sweet-spot model — great quality-to-size ratio
FP16 ~16 tok/s
Runs great → Gemma 3 4B
Gemma 3 4B 2.8 GB
Google's 4B — strong reasoning and instruction following
FP16 ~11 tok/s
Runs well → StableLM 2 1.6B
StableLM 2 1.6B 1.0 GB
Stability AI's efficient small model
FP16 ~30 tok/s
Runs great → Gemma 3n E4B
Gemma 3n E4B 4.5 GB
Google's most capable on-device model — 4B effective, multimodal
Q8 ~14 tok/s
Runs well → Gemma 3 12B
Gemma 3 12B 7.3 GB
Google's largest open model — near-frontier quality
Q6 ~11 tok/s
Tight fit → Mistral Nemo 12B
Mistral Nemo 12B 7.1 GB
Mistral's 12B — Tekken tokenizer, 128K context
Q6 ~11 tok/s
Tight fit → Gemma 3n E2B
Gemma 3n E2B 3.0 GB
Google's on-device multimodal — 2B effective params with vision and audio
FP16 ~11 tok/s
Runs well →
Reasoning 7 models
🧠
DeepSeek R1 Distill 1.5B 1.1 GB
Distilled reasoning — chain-of-thought in a tiny package
FP16 ~30 tok/s
Runs great → 🧠
DeepSeek R1 Distill 8B 4.9 GB
Llama-based R1 distill — best open reasoning at 8B
Q8 ~12 tok/s
Runs well → 🧠
Qwen3 8B 5.2 GB
Alibaba's latest — thinking mode with strong reasoning
Q8 ~11 tok/s
Runs well → 🧠
DeepSeek R1 Distill 7B 4.7 GB
Distilled from DeepSeek R1 — strong step-by-step reasoning
Q8 ~12 tok/s
Runs well → Mistral 7B
Mistral 7B 4.3 GB
High-quality reasoning and analysis
Q8 ~12 tok/s
Runs well → Phi-4 Mini
Phi-4 Mini 8.7 GB
Microsoft's reasoning model — exceptional for its size
FP16 ~13 tok/s
Runs well → Phi-4 Medium
Phi-4 Medium 8.0 GB
Microsoft's 14B reasoning model — frontier-class performance
Q6 ~9 tok/s
Tight fit →
Code 6 models
Qwen 2.5 Coder 0.5B
Qwen 2.5 Coder 0.5B 477 MB
Tiny code completion model — autocomplete on any device
FP16 ~83 tok/s
Runs great → Qwen 2.5 Coder 1.5B
Qwen 2.5 Coder 1.5B 1.1 GB
Best small code model — great for autocomplete
FP16 ~31 tok/s
Runs great → Qwen 2.5 Coder 7B
Qwen 2.5 Coder 7B 4.4 GB
Best open 7B code model — rivals GPT-4 on coding benchmarks
Q8 ~12 tok/s
Runs well → Qwen 2.5 Coder 3B
Qwen 2.5 Coder 3B 2.0 GB
Strong code generation and editing at 3B
FP16 ~16 tok/s
Runs great → 👨‍💻
StarCoder2 3B 1.9 GB
BigCode's multilingual code model — 600+ languages
FP16 ~16 tok/s
Runs great → 👨‍💻
DeepSeek Coder 6.7B 3.8 GB
DeepSeek's code model — strong at generation and debugging
Q8 ~14 tok/s
Runs well →
Vision 4 models
👁️
SmolVLM 500M 490 MB
Tiny vision-language model — describe images on any device
FP16 ~91 tok/s
Runs great → Qwen 2.5 VL 7B
Qwen 2.5 VL 7B 4.6 GB
State-of-the-art vision-language — image, video, document understanding
Q8 ~12 tok/s
Runs well → 👁️
Moondream 2B 1.3 GB
Small but capable vision model — image Q&A and captioning
FP16 ~27 tok/s
Runs great → 👁️
LLaVA 1.6 7B 4.5 GB
Leading open vision model — image understanding and reasoning
Q8 ~13 tok/s
Runs well →
Embedding 4 models
🔢
Nomic Embed Text 137 MB
Open-source embedding with 8K context — long document search
FP16 ~200 tok/s
Runs great → 🔢
GTE Large 335 MB
High-quality embeddings — top of MTEB benchmark at its size
FP16 ~149 tok/s
Runs great → 🔢
all-MiniLM-L6-v2 23 MB
Fast sentence embeddings — ideal for semantic search
FP16 ~200 tok/s
Runs great → 🔢
BGE Small 34 MB
Compact BAAI embedding — great for RAG pipelines
FP16 ~200 tok/s
Runs great →
Image 4 models
🎨
Stable Diffusion Turbo 2.2 GB
1-step image generation — instant results
FP16 ~48 tok/s
Runs great → 🎨
Stable Diffusion 3.5 Medium 4.6 GB
Latest SD architecture — excellent image quality
FP16 ~22 tok/s
Runs great → 🎨
SDXL Turbo 6.5 GB
High-res 1-step generation — 1024×1024 in seconds
FP16 ~15 tok/s
Runs great → 🎨
FLUX.1 Schnell 12.1 GB
Black Forest Labs' fast model — stunning quality in 4 steps
Q6 ~11 tok/s
Tight fit →
Voice 4 models
🗣️
Kokoro 82M 183 MB
24 natural voices — instant TTS
FP16 ~200 tok/s
Runs great → 🗣️
OuteTTS 0.3 500M 500 MB
Voice cloning and natural TTS — zero-shot voice synthesis
FP16 ~100 tok/s
Runs great → 🗣️
KittenTTS Nano 19 MB
Tiny, fast voice synthesis
FP16 ~200 tok/s
Runs great → 🗣️
Dia 1.6B 1.6 GB
Nari Labs dialogue TTS — multi-speaker with emotion
FP16 ~30 tok/s
Runs great →
Transcription 6 models
Whisper Medium
Whisper Medium 1.5 GB
Strong multilingual transcription
FP16 ~67 tok/s
Runs great → 🎙️
Distil-Whisper Large V3 1.5 GB
6x faster than Whisper Large — nearly same accuracy
FP16 ~67 tok/s
Runs great → Wav2Vec2 Base
Wav2Vec2 Base 231 MB
High-accuracy English ASR
FP16 ~200 tok/s
Runs great → Whisper Small
Whisper Small 488 MB
Good accuracy-speed balance — 99 languages
FP16 ~200 tok/s
Runs great → Whisper Tiny
Whisper Tiny 89 MB
Real-time speech recognition — runs on anything
FP16 ~200 tok/s
Runs great → Whisper Large V3
Whisper Large V3 3.1 GB
Best open transcription — near-human accuracy
FP16 ~32 tok/s
Runs great →
Too heavy 1 models
Llama 4 Scout
Llama 4 Scout 63 GB
Meta's MoE model — 109B total params, 17B active per token
needs 42.0 GB
Too heavy →
BROWSE OTHER TIERS
Low tier
4 GB
Mid tier
8 GB
Ultra tier
32 GB