Tencent’s Yuanbao search agent uses hybrid OPD to keep general skills
_reachsumit · x · 2026-08-04
Tencent presents Cross-Domain Hybrid OPD, a training framework for its Yuanbao search agent that tries to avoid the usual tradeoff between search specialization and general capability.
- The system combines agentic RL for autonomous search with cross-domain expert on-policy distillation.
- The goal is to improve search behavior without causing an “alignment tax” that degrades general-purpose skills.
- Built on the Hunyuan3 architecture, the report says the resulting model keeps competitive search performance while also improving broad capabilities across domains.
More from Companies & People
- Zuckerberg: Every Business Will Have Its Own Customized AI in the Future — rohanpaul_ai · 2026-08-04
- MiniMax Exec Reviews Hailuo AI's Growth, Announces Push for Open Source — VoidAsuka · 2026-08-04
- Jensen Huang: Closed Models Are Actually Cheaper; Open Source Is About Control — rohanpaul_ai · 2026-08-04
- OpenAI appears to publicly call out Apple in a company post titled “Apple is getting this wrong” — apples_jimmy · 2026-08-04
- Numerai forms a team to build autonomous AI for quantitative research — AndrewCritchPhD · 2026-08-04
- OpenAI says Apple emailed the wrong person in its lawsuit over file access — ChrisGPT · 2026-08-04