Tencent Hunyuan Hy3 officially released: significant gains in agents and product experience
Original title:腾讯混元 Hy3 发布:Agent 能力和产品体验跃升
AI summary
Tencent Hunyuan released Hy3, with significant progress on reasoning, agent and long-context tasks, comparable to flagship models with 2-5 times the parameter scale. Its internal blind-test average score is 2.67/4, better than GLM5.1 (2.51/4). The hallucination rate fell from 12.5% to 5.4%, the common-sense error rate from 25.4% to 12.7%, the multi-turn problem rate to 7.9%, and MRCR rose from 42.9% to 75.1%. Task resolution rate jumped from 72% to 90%, and average time fell 34%. API pricing is 1 yuan for input and 4 yuan for output per million tokens, and 0.25 yuan for cache hits. It has been open-sourced under the Apache 2.0 license on GitHub, HuggingFace and other platforms.
Key Takeaways
- Tencent Hunyuan released Hy3,with significant progress on reasoning,agent and long-context…
- Its internal blind-test average score is 2.
- 67/4,better than GLM5.
- The hallucination rate fell from 12.
- 4%,the common-sense error rate from 25.
Quick Facts
- Published
- Updated
- Source
- 公众号:腾讯混元
- Category
- Model Releases
- AI score
- 70
腾讯混元 Hy3 不堆参数,靠 agent 能力和产品实测说话,幻觉率减半、多轮问题率大降,开源+降价让开发者能直接上手,是国内大模型里少见的产品导向发布。
Related entities
Source & copyright
This article is based on curated titles and AI summaries from the AIHOT public API, with attribution and links back. Original copyright remains with the original publishers and authors; this site does not mirror full texts.
How to Cite
When citing this article, keep the title, date, and permalink, and credit Orange News as the source.
Orange News: Tencent Hunyuan Hy3 officially released: significant gains in agents and product experience. July 6, 2026. https://new.nat6.net/en/news/cmr8w4ho5013isllsr07z577p/