GLM-4.7 Flash 18 GB
Zhipu's 30B-A3B MoE — 3B active, efficient local coding & agents, 59.2% SWE-bench Verified, 128K context
MIN TIER
Ultra
SMALLEST QUANT
Q2 · 11.0 GB
CONTEXT WINDOW
128K tokens
PARAMETERS
3B
QUANTIZATION OPTIONS
| Quant | File size | Quality | |
|---|---|---|---|
| FP16 | 60.0 GB | 100% | Best quality |
| Q8 | 32.0 GB | 95% | |
| Q6 | 23.0 GB | 85% | |
| Q5 | 20.0 GB | 78% | |
| Q4 | 18.0 GB | 70% | |
| Q3 | 14.0 GB | 58% | |
| Q2 | 11.0 GB | 42% |
DEVICE COMPATIBILITY
Runs on these devices (53)
🖥️
Gaming PC (RTX 5090) 64GB
Windows Q6 · ~78 tok/s · 23.0 GB
Runs well 🖥️ Mac Pro M2 Ultra 192GB
macOS FP16 · ~13 tok/s · 60.0 GB
Runs great 🖥️ Gaming PC (RTX 4090) 64GB
Windows Q5 · ~50 tok/s · 20.0 GB
Tight fit 🖥️ Mac Studio M1 Ultra 64GB
macOS Q8 · ~25 tok/s · 32.0 GB
Runs well 🖥️ Mac Studio M2 Ultra 64GB
macOS Q8 · ~25 tok/s · 32.0 GB
Runs well 🖥️ Gaming PC (RTX 3090) 64GB
Windows Q5 · ~47 tok/s · 20.0 GB
Tight fit 🖥️ Gaming PC (RX 7900 XTX) 64GB
Windows Q5 · ~48 tok/s · 20.0 GB
Tight fit 🖥️ Gaming PC (RTX 4080) 32GB
Windows Q3 · ~51 tok/s · 14.0 GB
Tight fit 🖥️ Gaming PC (RTX 5080) 32GB
Windows Q3 · ~69 tok/s · 14.0 GB
Tight fit 🖥️ Mac Studio M4 Max 64GB
macOS Q8 · ~17 tok/s · 32.0 GB
Runs well 🖥️ Gaming PC (RTX 3080) 32GB
Windows Q4 · ~42 tok/s · 18.0 GB
Tight fit 🖥️ Gaming PC (RX 7800 XT) 32GB
Windows Q3 · ~45 tok/s · 14.0 GB
Tight fit 💻 MacBook Pro M1 Max 64GB
macOS Q8 · ~13 tok/s · 32.0 GB
Runs well 💻 MacBook Pro M2 Max 64GB
macOS Q8 · ~13 tok/s · 32.0 GB
Runs well 🖥️ Gaming PC (RTX 5070) 32GB
Windows Q5 · ~34 tok/s · 20.0 GB
Tight fit 🖥️ Gaming PC (Arc A770) 32GB
Windows Q3 · ~40 tok/s · 14.0 GB
Tight fit 🤖 Atom 1 64GB
Linux Q8 · ~9 tok/s · 32.0 GB
Runs well 🤖 Atom 1 128GB
Linux FP16 · ~5 tok/s · 60.0 GB
Runs well 🖥️ Gaming PC (RTX 4070) 32GB
Windows Q5 · ~25 tok/s · 20.0 GB
Tight fit 💻 MacBook Pro M5 Max 48GB
macOS Q8 · ~19 tok/s · 32.0 GB
Tight fit 🖥️ Gaming PC (RTX 3070) 32GB
Windows Q3 · ~32 tok/s · 14.0 GB
Tight fit 💻 MacBook Pro M4 Max 48GB
macOS Q8 · ~17 tok/s · 32.0 GB
Tight fit 🖥️ Gaming PC (Arc B580) 32GB
Windows Q5 · ~23 tok/s · 20.0 GB
Tight fit 🖥️ Mac Studio M3 Ultra 96GB
macOS FP16 · ~14 tok/s · 60.0 GB
Tight fit 💻 MacBook Pro M1 Max 32GB
macOS Q6 · ~17 tok/s · 23.0 GB
Tight fit 💻 MacBook Pro M2 Max 32GB
macOS Q6 · ~17 tok/s · 23.0 GB
Tight fit 💻 MacBook Pro M3 Max 36GB
macOS Q6 · ~17 tok/s · 23.0 GB
Tight fit 🖥️ Gaming PC (RTX 3060) 32GB
Windows Q5 · ~18 tok/s · 20.0 GB
Tight fit 💻 MacBook Pro M3 Max 96GB
macOS FP16 · ~7 tok/s · 60.0 GB
Tight fit 🖥️ Mac Mini M4 Pro 48GB
macOS Q8 · ~9 tok/s · 32.0 GB
Tight fit 💻 MacBook Pro M4 Pro 36GB
macOS Q6 · ~12 tok/s · 23.0 GB
Tight fit 🖥️ Mac Mini M4 Pro 24GB
macOS Q3 · ~20 tok/s · 14.0 GB
Tight fit 💻 MacBook Pro M5 Pro 24GB
macOS Q3 · ~21 tok/s · 14.0 GB
Tight fit 💻 MacBook Pro M1 Pro 32GB
macOS Q6 · ~9 tok/s · 23.0 GB
Tight fit 💻 MacBook Pro M2 Pro 32GB
macOS Q6 · ~9 tok/s · 23.0 GB
Tight fit 🤖 Atom 1 32GB
Linux Q6 · ~9 tok/s · 23.0 GB
Tight fit 🖥️ Mac Mini M2 Pro 32GB
macOS Q6 · ~9 tok/s · 23.0 GB
Tight fit 🖥️ Gaming PC (RTX 4060) 32GB
Windows Q3 · ~19 tok/s · 14.0 GB
Tight fit 💻 MacBook Pro M3 Pro 36GB
macOS Q6 · ~7 tok/s · 23.0 GB
Tight fit 🖥️ Mac Mini M4 32GB
macOS Q6 · ~5 tok/s · 23.0 GB
Tight fit 💻 MacBook Pro M1 Pro 16GB
macOS Q2 · ~18 tok/s · 11.0 GB
Tight fit 💻 MacBook Pro M2 Pro 16GB
macOS Q2 · ~18 tok/s · 11.0 GB
Tight fit 🖥️ Mac Mini M2 Pro 16GB
macOS Q2 · ~18 tok/s · 11.0 GB
Tight fit 💻 MacBook Pro M3 Pro 18GB
macOS Q2 · ~14 tok/s · 11.0 GB
Tight fit 💻 MacBook Pro M5 16GB
macOS Q2 · ~14 tok/s · 11.0 GB
Tight fit 💻 MacBook Air M4 16GB
macOS Q2 · ~11 tok/s · 11.0 GB
Tight fit 🖥️ Mac Mini M4 16GB
macOS Q2 · ~11 tok/s · 11.0 GB
Tight fit 💻 MacBook Air M3 16GB
macOS Q2 · ~9 tok/s · 11.0 GB
Tight fit 💻 MacBook Air M2 16GB
macOS Q2 · ~9 tok/s · 11.0 GB
Tight fit 🎮 Steam Deck OLED 16GB
Linux Q2 · ~8 tok/s · 11.0 GB
Tight fit 💻 MacBook Air M1 16GB
macOS Q2 · ~6 tok/s · 11.0 GB
Tight fit 💻 MacBook Pro M1 16GB
macOS Q2 · ~6 tok/s · 11.0 GB
Tight fit 🖥️ Mac Mini M1 16GB
macOS Q2 · ~6 tok/s · 11.0 GB
Tight fit Too heavy for these devices (14)
💻
MacBook Air M2 8GB
macOS needs 11.0 GB
Too heavy 💻 MacBook Air M1 8GB
macOS needs 11.0 GB
Too heavy 📱 iPhone 16 Pro 8GB
iOS needs 11.0 GB
Too heavy 📱 iPhone 15 6GB
iOS needs 11.0 GB
Too heavy 📱 Galaxy S25 Ultra 12GB
Android needs 11.0 GB
Too heavy 📱 Galaxy S24 8GB
Android needs 11.0 GB
Too heavy 📱 Pixel 9 Pro 16GB
Android needs 11.0 GB
Too heavy 📱 iPad Pro M4 16GB
iOS needs 11.0 GB
Too heavy 🖥️ Mac Mini M1 8GB
macOS needs 11.0 GB
Too heavy 🖥️ Mac Mini M2 8GB
macOS needs 11.0 GB
Too heavy 💻 Snapdragon X Elite Laptop 16GB
Windows needs 11.0 GB
Too heavy 📱 OnePlus 13 16GB
Android needs 11.0 GB
Too heavy 🍓 Raspberry Pi 5 8GB
Linux needs 11.0 GB
Too heavy 💻 MacBook Air M3 8GB
macOS needs 11.0 GB
Too heavy RUN WITH THESE APPS
Beginner Free
LM Studio
Beautiful desktop app for running local AI models
Tinkerer Free
Ollama
Lightweight CLI tool for running models in the background
Beginner Free
Jan
Open-source ChatGPT-like desktop app
Expert Free
opencode
Open-source terminal coding agent — bring your own local model
Tinkerer Free
Cline
Autonomous coding agent for VS Code with local model support
Tinkerer Free
Continue
Open-source autocomplete and chat for VS Code and JetBrains
Expert Free
Aider
Terminal AI pair programming, git-first
HOW TO RUN
Step-by-step run guides coming soon — check the apps above to get started today.