Subscribe
Share
Share
Embed
Daily AI news and research, distilled. UpNext AI breaks down the most important developments in artificial intelligence—from major industry moves to cutting-edge papers.
Welcome to the UpNext AI podcast. It's Friday, July 10th, 2026, and here's what matters in AI today.
Our top story: OpenAI says GPT-5.6 is now the preferred model for Microsoft 365 Copilot. TechCrunch reports the announcement comes after recent chatter that Microsoft had been using more of its own in-house models in some apps to cut costs. OpenAI says GPT-5.6 will support Microsoft users across its productivity suite, including Word, Excel, PowerPoint, and Cowork. The exact meaning of 'preferred model' is still not fully clear, but the practical signal is clear enough: OpenAI is saying its newest model family remains central inside Microsoft’s workplace AI stack.
That leads directly to the broader launch itself. TechCrunch reports that OpenAI unveiled the GPT-5.6 family in three variants: Sol, Terra, and Luna, and positioned them for enterprise work, coding, and scientific research. OpenAI also describes GPT-5.6 as its strongest cybersecurity model yet, with support for defensive tasks like threat modeling, code review and patching, and blue teaming. The company says the models are available across ChatGPT, Codex, and the OpenAI API. TechCrunch also reports pricing for the top two tiers at 5 dollars input and 30 dollars output for Sol, and 2 dollars 50 input and 15 dollars output for Terra. So this looks less like a narrow model refresh and more like OpenAI trying to plant a platform-wide flag.
For today’s research note, a new arXiv paper describes a clinical-reasoning model called HCC-STAR for hepatocellular carcinoma, or HCC, the most common form of liver cancer. The researchers say the model reads routine medical-record narratives and produces risk-based staging, ranked treatment suggestions that follow guidelines, and individualized survival estimates. In a multi-center cohort of 6,668 patients from 12 hospitals in China, the paper reports a hypothetical median overall survival of 51 months under adherence to HCC-STAR recommendations, compared with 29 and 32 months under the BCLC and CNLC approaches cited in the study. Risk stratification here just means sorting patients into more precise outcome groups so treatment can better match the case. Bottom line: it’s one paper, not a new standard of care, but it shows how medical LLMs are being pushed toward structured decision support with measurable clinical comparisons.
...Are you building apps with voice? Elevate your app's voice capabilities with ElevenLabs. Their API is a game changer for embedding dynamic, responsive voice interactions in your applications, providing unprecedented realism, flexibility and latency. In fact, you're listening to one of their voices - right - now. If you are a developer looking to elevate user experience with natural voice interfaces, this is your solution. Visit up next dot fm slash eleven to check out their latest offerings. ...
First, Google will now tell users when an ad was created or edited with AI. The Verge reports the label will appear in My Ad Center under a 'how this ad was made' tab for ads on Google Search, Google Discover, and YouTube. Ads made with Google’s own generative ad tools will be labeled automatically.
Next, Meta is unwinding its 2 billion dollar Manus acquisition. The Financial Times reports that Tencent is leading a deal to reverse Meta’s takeover of the AI agent startup after Beijing ordered the move, with Tencent set to become the largest shareholder.
And finally, OpenAI is also pushing harder into longer-running agent workflows. Ars Technica reports that its rebranded Codex effort, through ChatGPT Work, is designed for independent workflows that can run for hours if needed, while still asking users to approve important actions.
Before we wrap up, a quick note: this podcast is generated with the assistance of AI and is intended for informational purposes only. All referenced articles, research, and commentary remain the property of their original authors and publishers.
If you enjoyed this episode, don't forget to subscribe, rate, and leave us a review! And that's your briefing for today. Full source links are in the episode notes, and we'll be back Monday with what's up next!