🚨 AI News | TestingCatalog

GOOGLE 🔥: Lyria 3.5 has been released and is now available on Google AI Studio, Gemini app and APIs.

> "Our advanced, full-song generative model with enhanced audio fidelity and vocal clarity, optimized for precise prompt adherence and rich, cohesive musical arrangements."

Claude Code on iOS will get support for / commands as well.

Does anybody have it already? 👀

DAILY AI BRIEF 🗞 — Sept 4

OPENAI 🔥:
> GPT-6 Astra is officially out — SOTA across computer use, coding, science, and cyber. 98% FrontierMath Tier 4, 63% / 99.9% on ARC-AGI-3 (standard vs adapter harness), 100% ExploitBench.
> Not in your ChatGPT dropdown yet. Rolling to a limited set of orgs / Daybreak first; Plus, Pro, Business, Enterprise, API, and AWS “over the coming days.” Hope is weekend. Until then, enjoy the banked limit resets.

MICROSOFT 🔥:
> MAI-Transcribe-2 is live. Claims fastest + cheapest + most accurate speech-to-text: #1 on FLEURS across 60 languages (5.2% avg WER), 10× GPT-Transcribe, 7× Scribe v2, 5× Gemini 3.5 Transcribe. Foundry + MAI Playground + OpenRouter.

GOOGLE 🔥: (for Ultra Trusted Testers?)
> Gemini Live can now use Workspace connectors — Gmail, Keep, Docs. Conversational inbox search, voice-first drafting, brain-dump-to-lists.
> New Labs experiment: Play with Putty — real-time collaborative vibe coding. Multiple people in one workspace building sites/tools live. Waitlist.

TESLA 🔥:
> Cybercab is in the Austin robotaxi fleet. No steering wheel, no pedals, butterfly doors, public rides starting in limited zones. Elon: the future is here.

NVIDIA 🔥:
> Confirmed $12.9B acquisition of Hugging Face. Jensen: platform stays open, multi-cloud, multi-accelerator, no NVIDIA compute required. 18M developers, 3M+ models. Largest Nvidia buy ever. Close targeted for 2027, subject to regulatory approval.

* Used Grok to compose this brief, cherry-picking the news and doing some post-editing.

OpenAI announced Defense Factory, a new continuous defense concept.

"Defender's window" is the gap between the cyber capabilities of open-weight models and the current security level.

Without continuous defense practices that rely on frontier models, this gap will close, enabling malicious actors to freely abuse vulnerable systems.

> "Agents can now conduct long-running cyber operations by abusing increasingly available open-weight models. In response, at OpenAI we’re building a Defense Factory. An automated defense operation to find, validate, and fix vulnerabilities continuously."

> "Defenders have two structural advantages. They can give agents direct access to their code and use frontier models to get a head start over attackers abusing broadly available open-weight models."

What does it mean?

“Everything that makes it successful is exactly what makes it dangerous.”

Internet won't be safe anymore 🤖

OPENAI 🔥: ChatGPT paid users will get one banked reset for every day they don’t have Astra access, starting today.

Sam Altman - i am hopeful that you can use it this weekend! but can’t promise yet.

I hope we can get at least 3-5 resets.

Astralogy 🔮

Google announced a new Google Labs experiment called "Play with Putty", available via a waitlist for US users.

> Play with Putty is a collaborative vibe coding tool that lets you build tools and websites together in real time.

OPENAI 🔥: Astra scored 63% on ARC-AGI-3 with a standard harness.

> GPT-6 Astra surpasses the human baseline in action efficiency on ARC-AGI-3. It used fewer actions than the median tested human on 96% of levels.

> A key behavior observed in GPT-6 Astra was its ability to turn unfamiliar environments into compact symbolic world models. It represented game mechanics as logical rules and developed its own domain-specific language shorthand to track state and plan actions.

Stunning 👀👀👀

BREAKING 🔥: GPT-6 Astra has been officially announced and will roll out in the coming days.

> Astra was pre-trained on more than 100,000 GPUs at Stargate, and GPT-5.6 Sol has been helping supervise its training.

> "Across the plans, Astra will be included in the normal usage allocation, and you will be able to use 100% of it for Astra."

AGI? 👀

OPENAI 🔥: GPT-6 Astra scored 98.6% on ARC-AGI-3 and topped most of the other benchmarks.

74.% on DeepSWE 1.1, big leaps on FrontierMath Tier 4, Terminal-Bench Science 0.1 and more.

OPENAI 🔥: GPT-6 Astra will initially be available only to organizations in its Daybreak Access program.

AGI today, but only in Daybreak 👀

Over the next several days, it will be released to all Plus, Pro, Business, and Enterprise users.

GOOGLE 🔥: Gemini Live can now use Google Workspace connectors, including Gmail, Keep, and Docs.

That's big; I was waiting for that for a while!

BREAKING 🔥: ASTRA, AKA GPT-6 WILL BE RELEASED TODAY!

A lot of testing time today 👀

OPENAI 🔥: The next upgrade for the image model is being prepared for the release.

“Image creation got a major upgrade" "Higher-quality results, faster generation, and smarter creative tools"

Earlier, OpenAI has been spotted testing new image model on Image Arena as well.

GPT-Image-2.5? 👀

DAILY AI BRIEF 🗞 — Sept 3

GOOGLE 🔥:
> Gemini 3.8 Flash is live — third Flash in six weeks. Workhorse for coding and agents, 1M context. Intro price $0.75 / $3.75 through Dec 31.
> Flash Cyber is the defender twin (vuln find + auto-patch). Gated to Fairwind: governments, infra, trusted maintainers.

META 🔥:
> Muse Spark 1.3 is out in Muse Code and the Meta Model API. Zuck: biggest jump yet on coding and agents, “almost too cheap to meter.”
> Beat GPT-5.6 and Opus 5 on DeepSWE 1.1. Watermelon 🍉 and Spark open weights are next.

OPENAI 🔥:
> gpt-6-astra slug is on the APIs. Employees and testers are posting about a possible drop today. Still not public.
> Already tagged Critical for cyber under the Preparedness Framework — extra safeguards, limited tools first.

XAI 🔥:
> Grok Bot is live on Android. Elon: “Grok @Bot now on Android.” Play Store, same “give it real work” agent as iOS.

* Didn't plan to post it initially cuz all the news was covered already, but here you go :P