Choose theme
AI Resources
LFM2.5-350M
LFM2.5-350M is Liquid AI's 350-million-parameter text model for compact local and edge applications.
Liquid AI recommends it for extraction, structured outputs and tool use, and does not recommend it for knowledge-intensive work or programming. Use this as a first read, not a recommendation. Open the original project before trusting details like terms, limits, privacy, cost, setup, or safety.
What it is
Compact instruction-tuned model
It supplies text generation to an application; it is not a ready-made assistant.
Why it stands out
Several deployment formats
Native weights, GGUF, ONNX, MLX and OpenVINO exports address different inference runtimes and hardware.
Availability
Model card and exports
Liquid AI's model page links the weights, formats and usage examples.
Why it matters
What makes it useful
A developer extracting fields from short records can use the model's documented structured-output path inside an application, then check the results against the records.
What to know
Where it fits
An inference runtime loads the selected export. The surrounding application supplies data and executes any requested tools; the model does not run those actions by itself.
Notable points
What stands out
The instruction-tuned model and the Base model serve different jobs. Liquid AI labels Base for fine-tuning, so the two downloads are not interchangeable choices for a conversational integration.
Before using
What to review
Choose an export supported by your runtime and device.
Check extraction results and tool requests; a small model or structured format does not establish correctness.
Reader fit
Who may find it relevant
Developers working on bounded text-processing tasks with constrained deployment targets.
It needs an inference setup and application integration; Liquid AI's stated limits matter for broader knowledge or coding tasks.
Editorial note
Why LifeHubber lists it
Its default tool calls use a Python-style list. Liquid AI also documents requesting JSON calls instead. An application expecting JSON needs to set that instruction and handle the returned format, rather than assuming every tool-use model speaks the same syntax.
Source links
Source materials
Reader note
Before relying on this entry
LifeHubber lists entries to help readers inspect AI projects, not to endorse them or prove they are safe, suitable, accurate, maintained, or right for a specific use. We do not verify every entry in depth. Before relying on anything listed, review the original materials, terms, privacy practices, limits, and risks that matter for your situation.
What to explore next
Plan the rest of a local model setup.
A compact model is one component. Continue with the runtime, hardware and data-path choices that determine how a local AI workflow actually runs.
More in AI Models
Keep browsing this category
Explore more AI model resources.
Gemma 4
google/gemma-4
A Google DeepMind Gemma 4 model family collection with public checkpoints including Gemma 4 12B, a dense multimodal model Google describes around local agentic workflows, native audio input, and encoder-free vision/audio handling.
MiniMax H3 Integrations
MiniMax-AI/awesome-minimax-h3-integration
A community-maintained MiniMax H3 integration index that maps checkpoints, hardware and VRAM starting points, runtimes, ComfyUI nodes, prompting tools, acceleration routes, and deployment options.
GLM-OCR
zai-org/GLM-OCR
A 0.9B multimodal OCR model for complex document parsing with several self-hosted or API deployment paths.
For project maintainers
Listed here? You can use the badge.
If you maintain a project with a current LifeHubber listing, you may add the optional “Listed on LifeHubber AI Resources” badge to its README, docs, or website. No introduction or permission request is needed.
