| | --- |
| | license: apache-2.0 |
| | language: |
| | - en |
| | - zh |
| | tags: |
| | - zen |
| | - zenlm |
| | - hanzo-ai |
| | - zen-2 |
| | pipeline_tag: text-generation |
| | library_name: transformers |
| | --- |
| | |
| | # Zen 2 72B — Instruct |
| |
|
| | > **Originally released**: 2024-06-08 |
| | > **Parameters**: 72B | **Architecture**: Zen 2 Architecture | **Context**: 128K | **License**: Apache 2.0 |
| |
|
| | Second generation of the Zen LM family (Q2 2024). Key breakthrough: **128K context window** via improved positional encoding and sliding window attention. |
| |
|
| | Instruction-tuned via RLHF and supervised fine-tuning. |
| |
|
| | ## Research Lineage |
| |
|
| | | Generation | Release | Architecture | Key Advancement | |
| | |------------|---------|--------------|-----------------| |
| | | Zen 1 | Q4 2023 | Zen 1 Architecture | Foundation, dense transformer, 4K context | |
| | | Zen 2 | Q2 2024 | Zen 2 Architecture | 128K context window | |
| | | Zen 3 | Q3–Q4 2024 | Zen 3 Architecture | Sparse MoE routing, multimodal | |
| | | Zen 4 | Q1 2025 | Zen 4 Architecture | 256K–1M context, extended families | |
| | | Zen 5 | 2025+ | Zen MoDE | Next-generation frontier | |
| |
|
| |
|
| | --- |
| | ## The Zen LM Family |
| |
|
| | Joint research collaboration: |
| | - **Hanzo AI** (Techstars '17) — AI infrastructure, API gateway, inference optimization |
| | - **Zoo Labs Foundation** (501c3) — Open AI research, ZIPs governance, decentralized training |
| | - **Lux Partners Limited** — Compute coordination and settlement layer |
| |
|
| | All weights Apache 2.0. Download, run locally, fine-tune, deploy commercially. |
| |
|
| | [HuggingFace](https://huggingface.co/zenlm) · [Chat free](https://hanzo.chat) · [API](https://api.hanzo.ai) · [Docs](https://zenlm.org) |
| |
|
