Tri-7B-Search-preview

A search-enabled LLM that combines efficient language modeling with real-time internet search capabilities.

Model Details

  • Base Model: Tri-7B
  • Parameters: 7B
  • Architecture: Transformer-based with integrated search functionality
  • License: Apache 2.0
  • Training: SFT with Qwen3-32B trajectories.

Demo & Implementation

Try it out with tri.oo.ai - a local, open-source implementation featuring:

  • Chat interface inspired by oo.ai
  • DuckDuckGo search integration
  • vLLM support for efficient inference
  • OpenAI-compatible API
Downloads last month
20
Safetensors
Model size
7.53B params
Tensor type
F32
ยท
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for trillionlabs/Tri-7B-Search-preview

Finetuned
(1)
this model
Quantizations
2 models