Brief, mostly facts
Agnes-3.0-Qwen: a 180B MoE multimodal model with 1M context
Released by Agnes AI with tool calling and text, image and video input. A GGUF quant is on Hugging Face under the Qwen Community License 1.0.
What the research found
Agnes AI released agnes-3.0-qwen, a ~180B hybrid-attention MoE multimodal reasoning model with 1M-token context, tool calling, and text/image/video input; bartowski published a GGUF quant on Hugging Face under the Qwen Community License 1.0.
Claims and how well the evidence supports them
- Agnes-3.0-Qwen has about 180B total parameters with 512 experts and 10 active per token 99%
- Context window is 1,048,576 tokens with maximum output 65,536 tokens 99%
- License is Qwen Community License 1.0 99%
- Model is distributed as GGUF quantization by bartowski on Hugging Face 94%
Evidence
- reddit.com https://reddit.com/r/LocalLLaMA/comments/1xaxz80/agnesaiagnes30qwen_qfn_finetune
- huggingface.co https://huggingface.co/bartowski/Agnes-AI_Agnes-3.0-Qwen-GGUF
Actionability rated medium.