Features
Mixture-of-Experts (MoE) architecture with 35B total, 3B active parameters
Agentic coding and tool calling capabilities
Multimodal reasoning (text + vision via encoder)
State-of-the-art mathematical reasoning
High throughput – comparable to dense 3B model speed
Apache 2.0 open-source license
Supports fine-tuning for custom tasks
Available via Hugging Face and GitHub
Docker-based deployment scripts
Compatible with Qwen API and local inference
Long context support (up to 32K tokens)
Multilingual support (English, Chinese, and others)
Optimized for consumer GPUs (e.g., RTX 4090)
Quantized versions (GGUF, AWQ) for efficient deployment
Local model execution for offline inference
Connect online APIs: OpenAI (ChatGPT), Anthropic (Claude), Google (Gemini), Meta (Llama), Mistral AI, Alibaba (Qwen),
123+ HuggingFace models available
Memory feature (coming soon) for context persistence
Minimalist UI with dark mode
Open source under MIT license on GitHub
Cross-platform desktop app (Mac, Windows, Linux)
Version 0.8.3 with bug fixes and performance improvements
