# Hy3: Self-Hosting Hardware, Scenarios and Commercial License

> Large private coding, productivity and tool-using agents that can support an official eight-accelerator tensor-parallel worker.

- Verified: 2026-09-12
- Released: 2026-07-06
- Canonical: https://chinaapi.ai/open-model-deployment/hy3/
- Back to directory: https://chinaapi.ai/open-model-deployment/
- Machine-readable dataset: https://chinaapi.ai/data/open-model-deployment.json

## Model facts

- Vendor: Tencent Hunyuan
- Parameters: 295B backbone plus 3.8B MTP layer total; 21B backbone per token active
- Native precision: BF16
- Context: 256K tokens
- Category: General and agentic foundation models
- Modalities: text → text
- Serving paths: Transformers, vLLM, SGLang
- Official weights: https://huggingface.co/tencent/Hy3?utm_source=chinaapi&utm_medium=research&utm_campaign=open-model-deployment
- Model documentation: https://github.com/Tencent-Hunyuan/Hy3?utm_source=chinaapi&utm_medium=research&utm_campaign=open-model-deployment

## Deployment evidence

- **Official launch / minimum — evidence C:** The vendor recommends eight H20-3e GPUs or accelerators with more memory for the documented vLLM and SGLang TP8 path. This proves an official launch shape, not an SLA. Source: https://github.com/Tencent-Hunyuan/Hy3?utm_source=chinaapi&utm_medium=research&utm_campaign=open-model-deployment
- **200-person team — evidence E:** Benchmark one TP8 worker against the 200-seat profile, then add an independent worker if maintenance and failure recovery must preserve service.
- **Commercial API — evidence C:** Official vLLM and SGLang OpenAI-compatible paths exist; a public service still needs redundant TP8 workers, admission control and pinned parsers. Source: https://github.com/Tencent-Hunyuan/Hy3?utm_source=chinaapi&utm_medium=research&utm_campaign=open-model-deployment

## Scenario evidence

- **Coding and productivity agents — Vendor-stated:** Tencent positions Hy3 for coding, productivity, reasoning and agent tasks. Published capability results are vendor evaluations, not ChinaAPI reproductions. Source: https://github.com/Tencent-Hunyuan/Hy3?utm_source=chinaapi&utm_medium=research&utm_campaign=open-model-deployment
- **Long-context tool workflows — Vendor-stated:** The official configuration supports 256K context and documents tool and reasoning parsers for vLLM and SGLang. Source: https://github.com/Tencent-Hunyuan/Hy3?utm_source=chinaapi&utm_medium=research&utm_campaign=open-model-deployment
- **Private OpenAI-compatible endpoint — ChinaAPI inference:** The official serving commands make an operator-controlled endpoint plausible, but capacity still depends on prompt length and concurrency.

## Commercial-use check

- License: Apache-2.0 — https://github.com/Tencent-Hunyuan/Hy3/blob/main/LICENSE?utm_source=chinaapi&utm_medium=research&utm_campaign=open-model-deployment
- MaaS / hosted service: No model-specific MaaS restriction identified in Apache-2.0.
- Attribution: Provide the license and required notices, preserve attribution notices, and mark modified files; trademark rights are not granted.

## Avoid or validate first

- Planning from the 21B active count instead of loading the full checkpoint
- Treating an eight-GPU launch command as a production capacity result

## Known limitations and open questions

- No ChinaAPI hardware or quality reproduction
- The official eight-GPU shape does not state concurrency or latency
- Full-context KV-cache demand must be measured separately

## Evidence boundary

Loading weights, completing a first forward pass and meeting a production latency/SLA are separate thresholds. The 200-person and commercial tiers require workload-specific measurement. License summaries are product research, not legal advice.

Generated by GENERATED BY scripts/gen_open_model_deployment.py.