A comparison of three leading open AI models from Chinese labs—Tencent's Hy4, DeepSeek's V4 Pro, and Alibaba's Qwen3.8 Max—evaluates their performance, pricing, and licensing for 2027. DeepSeek V4 Pro leads in verified coding benchmarks with an MIT license, making it a strong choice for accuracy and self-hosting. Tencent's Hy4 offers competitive pricing and agentic capabilities, particularly its Flash variant for cost-effective API use. Qwen3.8 Max is noted for its integration within the Alibaba ecosystem but comes with a more restrictive custom license and higher API costs. AI
IMPACT DeepSeek V4 Pro emerges as the top choice for coding tasks due to its verified performance and permissive license, while Tencent's Hy4 offers cost-effective solutions for general use and agentic applications.
RANK_REASON Comparison of multiple AI models based on benchmarks and pricing. [lever_c_demoted from research: ic=1 ai=1.0]
- Alibaba Group
- Claude Opus-5
- DeepSeek
- DeepSeek V4 Pro
- Hy4 Flash
- NeurAITool
- OpenRouter
- SWE-bench Verified
- Tencent
- Tencent Hy4
- TokenHub
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →