
EP 611 : Z AI’s GLM‑5.3‑Flash Revealed as Ox Alpha & OpenAI’s AGI Countdown – AI News Today
Last week the internet buzzed over a mysterious free model called Ox Alpha that seemed to out‑perform every contender. In today’s episode we finally crack the case: the model is Z AI’s brand‑new GLM‑5.3‑Flash, released with open weights and a price tag that’s a fraction of the frontier rivals. We break down why the cheap‑and‑powerful combo matters, how the entire launch ran on domestically‑made Chinese chips, and what the 57 Intelligence Index score tells us about the emerging price‑performance race.We then shift to OpenAI, where Sam Altman has put a deadline on AGI. A new TIME profile reveals that OpenAI expects an internal model meeting Altman’s “real‑AGI” bar by the end of the year, with research chief Mark Chen saying the lab is “80 % of the way” and Astra already able to conduct a week‑long research project solo. We discuss the implications of that claim, the recent July security breach post‑mortem, and how the push for safety may reshape frontier development.Our rapid‑fire roundup covers the hottest AI headlines: a startup building physics‑aware AI after turning down Jeff Bezos’s Project Prometheus, Instinct’s text‑only AI assistant that lives on a phone call, Bill Gates’ warning about an “AI tax,” and Anthropic opening Claude data to top universities. We also highlight new agentic models like Yutori’s Navigator n2 and Alibaba’s Qwen 3.8‑Flash‑Next preview.Finally, we surface the top AI tools that can supercharge your workflow this week—from Verse’s autonomous employee builder to Gridset’s on‑brand document creator, Talki’s multichannel support bot, and the latest flash models from Z AI, Fal, Google, and Alibaba.Tune in, share the episode with fellow AI enthusiasts, and hit subscribe so you never miss a breakthrough.Tools mentioned: Verse, Gridset, Talki, GLM‑3.5‑Flash, H3 Max, Gemini 3.5 Transcribe, Qwen3.8‑Flash‑Next.