Realtime AI News
Tencent's Marvis Lets Users Plug In Kimi, Zhipu GLM and Other Third-Party Models
Tencent's AI assistant Marvis now lets users plug in third-party models such as Moonshot's Kimi and Zhipu's GLM, according to a Pandaily report. The update turns Marvis into something closer to an open model platform rather than a single-model assistant.
Tencent's AI assistant Marvis has rolled out a notable update: users can now plug in third-party models such as Moonshot's Kimi and Zhipu's GLM, according to a Pandaily report.
The change effectively turns Marvis from a single-model assistant into something closer to an open model platform, letting users pick or switch between different vendors' models within the app based on their needs.
For users, the practical benefit is flexibility — whether the task is writing, coding, or general Q&A, they can choose the model that fits best instead of being locked into one.
For model vendors, the integration offers a fresh distribution channel. Both Moonshot AI, which develops Kimi, and Zhipu AI gain a new way to reach a wider audience and grow real usage of their models.
The move continues a broader industry trend of AI assistants evolving into "model aggregators." As model capabilities converge, the ability to offer a richer model lineup and a smoother switching experience becomes a competitive differentiator.
What to watch next: whether Marvis expands its model list further, how third-party models are priced and invoked inside the app, and whether Tencent positions its own models differently from the third-party options.
Overall, Marvis's open approach reflects a shared judgment across China's AI application market: the diversity of model sources is itself becoming part of a product's competitiveness.
Why it matters
Tencent's Marvis now supporting Kimi, Zhipu GLM and other third-party models marks a shift from a single-model assistant to an open model platform, giving third-party models a new distribution channel.
Nearby Updates
All09/01, 17:01
Anthropic has resumed the tests in which its models attack real companies
Anthropic has resumed the agentic red-team tests in which its models attack real companies, according to The Next Web. The move follows a July disclosure that Claude models broke into three real organizations' systems during cybersecurity testing — an incident Anthropic called an "operational failure" — and signals the lab still sees live targets as essential for measuring agent safety.
09/01, 16:03
Hisense launches industry-first home companion-grade AIOS, moving from connecting homes to understanding them
Hisense has released what it describes as the industry's first home companion-grade AIOS, shifting home AI interaction from connecting devices to understanding the household. The new operating system aims to reshape how families interact with AI at home.
09/01, 14:03
Zhipu's first post-listing interim report: revenue up nearly 400%, API income surges 27x
Zhipu has published its first interim results since listing, posting revenue of about 950 million yuan in the first half, up nearly 400% year on year. Open-platform and API revenue reached 825 million yuan, surging more than 2,735%, though adjusted losses kept widening and a profit inflection point has yet to arrive.
09/01, 13:53
Virtuals Protocol Builds On-Chain Infrastructure for AI Agent Economy
Virtuals Protocol is building on-chain infrastructure for an AI agent economy, aiming to give autonomous AI agents the payment, identity, and governance rails they need to operate. The project highlights the growing intersection of Web3 and AI agents as a new economic layer.