Side-by-side comparison of DeepSeek V3.2 (DeepSeek · China) and Hy-MT2-30B-A3B (Tencent) for self-hosted deployment of the open-weight model. DeepSeek V3.2 is rated conditional; Hy-MT2-30B-A3B is conditional. They part ways on licence: DeepSeek V3.2 is "MIT", Hy-MT2-30B-A3B is "Apache 2.0".
| Field | ||
|---|---|---|
| Summary | ||
| Verdict | Conditional 685B successor to V3 with DeepSeek Sparse Attention for long context, scalable RL for agentic tasks. Vendor claims parity with GPT-5 (Speciale variant exceeds). MIT licence keeps weights clean; Chinese-origin considerations unchanged. | Conditional Per the published Apache 2.0 LICENSE, the Hunyuan-MT2 30B-A3B weights are deployable without commercial restriction — this is a specialised fast-thinking translation model covering 36 languages, not a general-purpose chat backbone. The blockers for regulated EU use are Tencent's Chinese jurisdiction and a model card that names no training corpora; self-host the weights and prepare a deployer-side GPAI compliance file before processing personal data. |
| Last reviewed | 2026-04-15 | 2026-05-30 |
| Open-weight | ||
| Licence | MIT | Apache 2.0 |
| Commercial use | Yes | Unrestricted |
| Training data | Undisclosed | Undisclosed |
| Origin | China | China |
| Performance & pricing? | ||
| Quality index | 32/100 | — |
| Speed | 32 tok/s | — |
| Blended price | $0.32/M | — |
| Context window | — | — |
| Evidence | ||
| Sources | ||
No overlapping sources between the two entries.