Qwen3.5 Chat Playground

Load model in browser and verify chat in one page.

Model URL Quick Preset System Prompt Context Size (n_ctx) Batch Size (n_batch) Max Output Tokens Enable thinking mode

Idle

Tool demo: ask what time is it? or time now in Asia/Seoul
The page runs local JS tool get_current_time for verification.
You can toggle thinking mode before sending messages.
ONNX: paste a Hugging Face model repository URL or a direct .onnx file URL from that repository. Companion tokenizer, config, and graph files are loaded from the same repository.

Message