mirror of
https://github.com/hanzoai/zen.git
synced 2026-11-10 08:21:39 +00:00
2.9 KiB
2.9 KiB
Zen
Zen is Zen LM, the open model family from Zoo Labs Foundation, a 501(c)(3) non-profit. Open weights across language, code, vision, audio and retrieval: run them anywhere, or call them on the Hanzo API.
- Model Card: Zen 6 Architectural Specs & Benchmarks · Hugging Face Hub
- Family, model cards and papers: zenlm/zen · zenlm.org
- Weights: huggingface.co/zenlm
- Catalog and prices: hanzo.ai/models/zen
- Docs: docs.hanzo.ai/docs/models/zen
Model Lineup
| SKU | Context | Description |
|---|---|---|
zen5 (default) |
1M | Frontier open-weights ladder (glm-5.2 → deepseek-v4-pro) |
zen5-pro |
1M | High-capacity reasoning |
zen5-coder |
1M | Frontier code generation |
zen5-flash |
64K | Ultra-low latency open-weights |
zen5-mini |
32K | Lightweight local/edge deployment |
zen-vl |
128K | Vision and multimodal understanding |
zen-embedding |
8K | High-dimensional dense embeddings |
Use Zen through the Hanzo API
curl https://api.hanzo.ai/v1/chat/completions \
-H "Authorization: Bearer $HANZO_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model": "zen5", "messages": [{"role": "user", "content": "Hello"}]}'
The API speaks the OpenAI wire format, so any OpenAI SDK works with
base_url="https://api.hanzo.ai/v1". GET https://api.hanzo.ai/v1/models lists every Zen model
it serves.
Runnable Examples & Recipes
See examples/ for full runnable scripts and clients:
- Cloud API (
https://api.hanzo.ai/v1):- Python (streaming + reasoning tokens):
examples/cloud/client_cloud.py - TypeScript (native fetch):
examples/cloud/client_cloud.ts - Go (
net/http):examples/cloud/client_cloud.go - Rust (
reqwest):examples/cloud/client_cloud.rs - cURL script:
examples/cloud/client_cloud.sh
- Python (streaming + reasoning tokens):
- Local Execution (Offline Weights):
- Download weights:
examples/local/download_weights.sh - Run local inference server:
examples/local/run_local_server.sh - Python local client:
examples/local/client_local.py - Pure Rust local client:
examples/local/client_local.rs
- Download weights:
- Production Rust Recipes:
- Modular integration crate:
recipes/(ZenChatRecipe,ZenStreamAccumulatorfor reasoning tokens)
- Modular integration crate:
- AI Gateway & Custom Org Routing:
- Per-org on-prem routing:
examples/gateway/custom_org_routing.json
- Per-org on-prem routing: