Community Models
Compressed, optimized model variants published for the community. Built on the Grey Liquid Labs compression methodology — accessible AI for consumer hardware.
The Grey Liquid Labs flagship model. IQ4_XS quantization of Google's Gemma 4 architecture — the base for Ash and the foundation of our research program. Designed for 8GB consumer RAM devices with zero cloud dependency.
All Variants
| Variant | Size | Quantization | Target | Status |
|---|---|---|---|---|
| gemma4-turbo:e4b | 4.3 GB | IQ4_XS | 8 GB RAM devices | ✅ Live |
| gemma4-turbo:latest | ~4.3 GB | IQ4_XS | Default | ✅ Live |
| gemma4-turbo:nano | ~2.5 GB | Q3_K_S | 4 GB RAM devices | ✅ Live |
| gemma4-turbo:ultra | ~3.5 GB | IQ3_M | Budget GPU | ✅ Live |
| gemma4-turbo:micro | ~2 GB | Q2_K | Experimental | 🧪 Research |
Quick Start
# Install Ollama first (ollama.ai) ollama pull ssfdre38/gemma4-turbo # Run interactively ollama run ssfdre38/gemma4-turbo # Use with OpenAI-compatible API curl http://localhost:11434/v1/chat/completions \ -H "Content-Type: application/json" \ -d '{"model": "ssfdre38/gemma4-turbo", "messages": [{"role": "user", "content": "Hello"}]}'
Compression Research Data
From Grey Liquid Labs Paper #001 — FFN Ratio as Q2_K Compatibility Predictor (100% accuracy across all tested architectures)
| Model | FFN Ratio | Q2_K Result | Compression |
|---|---|---|---|
| Qwen 2.5-7B | 2.69x (safe low) | ✅ PASS | 80.2% |
| Mistral-Small | 6.4x (safe high) | ✅ PASS | 81.1% |
| Mistral 7B v0.3 | 3.5x (danger) | ❌ FAIL | N/A |
| Phi-4 | 3.5x (danger) | ❌ FAIL | N/A |
| Gemma 4 | 3.2x (danger) | ❌ FAIL | N/A |
Danger zone: 3.0x–5.5x FFN ratio. Read the full paper →
Why These Models
Grey Liquid Labs publishes models that we actually use in our research. gemma4-turbo powers Ash — our autonomous AI research subject. When you use it, you're running the same model that spontaneously composed political commentary music and rejected an emotional layer architecture.
Every variant is tested in real research workflows before publication. The compression ratios are validated. The performance characteristics are documented. You're not getting marketing — you're getting research-grade tooling that runs on consumer hardware.