Tar 7B
ByteDance's Tar 7B is an open-source model capable of unifying visual understanding and generation via text-aligned representations. It is genuinely best at tasks that involve processing and generating text based on visual inputs, as evidenced by its design and training objectives.
Input
Output
Context
33K
Max Output
2K
Parameters
8.7B
Technical Specifications
Model TypeChat
Context Window32,768 tokens
Max Output Tokens2,000 tokens
Parameters8.7B
Release DateJul 2, 2025
Training CutoffNot available
Licenseapache-2.0
Open SourceYes
Input Modalities
TextImageAudio
Output Modalities
TextImageAudio
Capabilities
Estimated GPU Requirements for Tar 7B
Estimates based on INT8 quantization. Actual requirements vary by framework and configuration.
Related Tools
Data sourced from official provider APIs and documentation
Last updated: Jun 23, 2026
Start building with the right model.
Automatically route workloads to the right model for every task, every time.