Run the Kimi K3 Architecture Locally with a 0.18B Test Model
A 0.18B-parameter Kimi K3 derivative that preserves the core layer types in a roughly 700 MB checkpoint for local architecture testing and development.
No approved media preview is available.
A 0.18B-parameter Kimi K3 derivative that preserves the core layer types in a roughly 700 MB checkpoint for local architecture testing and development.
Before real use, quickly verify source quality, freshness, and fit for your workflow.
2 link(s) checked without blocking signals.
Decision aid, not a safety guarantee.
What you get from it
Resource in AI & Automation with the public original source Kimi-K3-0.18B model card (huggingface.co). The full contents continue directly below.
Useful for AI & Automation, Resource, and adjacent workflows.
Open Kimi-K3-0.18B model card first and validate new tools or prompts in a test setup.
Even when visible for free, quality and fit still depend on the original source and your setup.
Before real use, quickly verify source quality, freshness, and fit for your workflow.
Read details
Kimi-K3-0.18B is a tiny, MIT-licensed derivative of Moonshot AI’s 2.8T-parameter Kimi K3. It reduces the model to four layers, eight experts and 0.18B total parameters, with roughly 0.10B activated per token. The single model.safetensors checkpoint is about 0.69 GB, making the architecture accessible on ordinary development hardware.
What it preserves
The miniature model retains Kimi K3’s main layer types: dense and MoE feed-forward blocks, KDA linear attention, and an MLA full-attention layer. That makes it useful for testing loaders, integrations, architecture support and compression workflows without downloading or serving the full 2.8T model.
Best use cases
- Validate Transformers compatibility and custom Kimi K3 model code
- Prototype inference, quantization or conversion workflows
- Inspect the MoE and attention implementation on limited hardware
- Build development tests around the Kimi K3 architecture
Loading caveat
The example loader requires trust_remote_code=True, which executes repository-provided Python code. Review the model repository and pin a revision before running it in a sensitive environment. The model is not currently hosted by a Hugging Face inference provider.
Discussion
Share practical experience, questions, or warnings with the community.
Sign in to join the discussion and vote on comments.
Sign in