Choose Your Deployment Paradigm
From edge-device inference to frontier-scale cloud, pick the architecture that matches your constraints.
Small & Local
Instant on-device execution with zero network dependency. Ideal for laptops, edge devices, and sensitive codebases.
Large & Local
Full-scale offline reasoning without recurring API fees or token throttle limits on dedicated workstations.
Open Cloud
Deploy unquantized open weights across high-concurrency cloud clusters and dedicated GPU instances.
Hybrid Routing
Filter PII locally, auto-delegate heavy reasoning to the cloud. Best of both worlds.
Open-Weight Model Directory
Verified open weights synchronized from 11 providers — DeepSeek, Meta, Qwen, Mistral, NVIDIA, and more. Filter by reasoning, coding, OCR, embeddings, and MoE architectures.
Latest Lab Note
Reproducible experiments, measured results, and the decisions behind them.
TamaBench — Small Models, Long Horizons
A lightweight benchmark that puts a small or local model in a persistent virtual-pet sandbox. The agent must plan across three simulated days, use structured tools, manage money and supplies, and recover when delayed consequences go wrong.
Quick Access
Jump to any section of the platform.