| Name | Modified | Size | Downloads / Week |
|---|---|---|---|
| Parent folder | |||
| textgen-portable-3.20-windows-cuda12.4.zip | 2025-12-08 | 935.7 MB | |
| textgen-portable-3.20-windows-vulkan.zip | 2025-12-08 | 216.4 MB | |
| textgen-portable-3.20-windows-cpu.zip | 2025-12-08 | 201.2 MB | |
| textgen-portable-3.20-linux-cuda12.4.zip | 2025-12-08 | 1.6 GB | |
| textgen-portable-3.20-linux-rocm.zip | 2025-12-08 | 627.3 MB | |
| textgen-portable-3.20-macos-x86_64.zip | 2025-12-08 | 166.4 MB | |
| textgen-portable-3.20-linux-vulkan.zip | 2025-12-08 | 303.6 MB | |
| textgen-portable-3.20-linux-cpu.zip | 2025-12-08 | 257.2 MB | |
| textgen-portable-3.20-macos-arm64.zip | 2025-12-08 | 199.4 MB | |
| README.md | 2025-12-07 | 1.9 kB | |
| v3.20 source code.tar.gz | 2025-12-07 | 24.9 MB | |
| v3.20 source code.zip | 2025-12-07 | 25.1 MB | |
| Totals: 12 Items | 4.6 GB | 11 | |
Image generation support!
Changes
- Image generation support: Generate images with
diffusersmodels like Z-Image-Turbo in a new "Image AI" tab. Features include: - 4bit/8bit quantization
torch.compilesupport- LLM-generated prompt variations
- PNG metadata for generation settings
- Gallery for past generations
- Progress bar
- OpenAI-compatible API endpoint for image generation
For a step-by-step tutorial, consult: Image Generation Tutorial
- Pass
bos_tokenandeos_tokento jinja2 templates, making it possible to use the template forSeed-OSS-36B-Instructand other models - Use
flash_attention_2by default for Transformers models
Bug fixes
- Fix API requests always returning the same
createdtime
Backend updates
- Update llama.cpp to https://github.com/ggml-org/llama.cpp/tree/0a540f9abd98915edb99fed47d80078ed8d2f343
- Update ExLlamaV3 to 0.0.17
Portable builds
Below you can find self-contained packages that work with GGUF models (llama.cpp) and require no installation! Just download the right version for your system, unzip, and run.
Which version to download:
- Windows/Linux:
- NVIDIA GPU: Use
cuda12.4. - AMD/Intel GPU: Use
vulkanbuilds. -
CPU only: Use
cpubuilds. -
Mac:
- Apple Silicon: Use
macos-arm64.
Updating a portable install:
- Download and unzip the latest version.
- Replace the
user_datafolder with the one in your existing install. All your settings and models will be moved.