Reporting AI nonsense. Latest AI News on AI Agents, Model Releases, Tools, Leaks, and Rumors 🗞️
Berlin, Germany
Joined October 2015
- Tweets31.7K
- Following998
- Followers81.5K
- Likes99.6K
OPENAI 🔥: Frontier AI training was already resumed on October 1st, according to the latest statement from OpenAI.
> Additionally, a substantial amount of compute will be consistently allocated to red teaming.
This is a very important part as the cost of aligned frontier AI development will go up proportionally to its intelligence. Eventually, this may impact token pricing as well.
Yet, even these measures might be not sufficient which would imply further cost increase.
“We leave in the era of underpriced compute.”
Over the past several months, we have substantially hardened the security of our research training environment. We’ve removed live internet access for all frontier models during training and built an offline world for them to use instead—a cached version of the web where they can still learn to browse, search, and use applications.
On September 20, after a model exploited a DNS misconfiguration to find a way outside of this offline world, we paused frontier training to strengthen containment further. We assembled teams of agents, powered by our most cyber-capable models, and tasked them with finding weaknesses in our research environments.
While they found no route to the internet they could easily exploit while concealing their activity from our monitors, they did find zero-day vulnerabilities, virtual machine escapes, and hardware side-channels for communication between agents. We used their discoveries to further strengthen our defenses and contribute fixes to open-source software.
After strengthening our containment architecture, we resumed our frontier RL training and inference on October 1. We’ve reserved substantial compute for continuous red-teaming, and made high-compute red-teaming part of the safety cases we use to decide whether frontier training and evaluation can proceed securely.
openai.com/index/securing-ou…
KIMI AI🔥: Kimi K3.1 Pro landed on the 8th spot on WebDev Arena, with Kimi AI moving to the 4th place among AI labs overall.
That a big move 👀
Big news: @Kimi_Moonshot is back in the top-4 labs in the Code Arena: WebDev with the release of Kimi K3.1 Pro!
This model just landed #8 in the Code Arena: WebDev with 1692 pts, matching Claude Opus-5 (Max) at 70% lower cost. This is a 9 place jump (+37 points) from Kimi K3 (Max) at #17 with 1655 pts.
Kimi K3.1 Pro improved from Kimi K3 across all categories. Highlights:
- #7 in Brand & Marketing (up from #9)
- #9 in Simulations (up from #20)
Cost efficiency is another upgrade. Priced at $2/$10 per million input/output tokens, Kimi K3.1 Pro (Max) has a median cost per task of $1.23 (29% less than Kimi K3). See its placement compared to the Pareto frontier in thread.
With open weights coming soon, Kimi K3.1 Pro would be the new #1 open model in the Code Arena: WebDev, overtaking DeepSeek-v4.1-Flash. Stay tuned!
See @Kimi_Moonshot’s ranking by lab, and 3D generations from Kimi K3.1 Pro below in thread.
Huge congrats to the @Kimi_Moonshot team on this release!
BREAKING 🔥: Kimi AI officially announced Kimi K3.1 Pro, a 2.8T total and 104B active params MoE open model.
On benchmarks, Kimi K3.1 Pro performs on par with leading AI models like Opus 5.5 and GPT-6 Astra, while being significantly cheaper.
Open weights are coming on November 4th as well!
Introducing Kimi K3.1 Pro: our most capable 3T-class open-weight model.
🔹 1M-token context, 2.8T total / 104B active params (MoE), Multimodal.
🔹 Full-project coding: games in Unreal, Unity & Godot.
🔹 Excels at parametric 3D modeling in Blender & CAD.
🔹 More reliable at finance work: Accurate modeling with hallucination down by 65%.
🔹 Matching the strongest proprietary models across coding, agentic and knowledge benchmarks.
Kimi K3.1 Pro and K3.1 Pro Highspeed (2–3× faster) are now live on Kimi.ai, Kimi Work, Kimi Code, and the Kimi API.
Open weights by Nov 4, 2026.
🔗 API: platform.kimi.ai
🔗 Tech blog: kimi.ai/blog/kimi-k3-1-pro
SPACEXSI 🔥: Cursor now has Computer Use capabilities and can operate other applications on your laptop.
The feature is especially usefull for app development, as it can also run mobile simulators.
Computer Use is activated on demand or via @-mentioning "Computer Use" command.
Cursor is a Superapp now! 👀
ByteDance released Seedance 2.5 4K, featuring richer visual detail.
> Seedance 2.5 4K offers richer color gradations and smoother transitions thanks to 10-bit color depth.
> Users can use Draft mode to experiment with low-cost 480p previews and then upscale them to 4K.
BREAKING 🔥: Kimi K3.1 Pro is now rolling out on Kimi AI web and mobile.
> Kimi K3.1 Pro - Complex tasks and professional work, all-around flagship.
Testing time! 👀
Early outputs from the community 👀
testingcatalog.com/kimi-k3-1…
Kling AI has officially released Kling 4.0, its next-generation video model built for more controlled, consistent video generation.
> Omni Reference combines up to 15 images, videos, subjects, or voices in a single task.
> Each generation runs up to 30 seconds and supports up to 10 keyframes for more controlled scene progression.
Kling 4.0 gives creators more control over how their scenes unfold - from complex environments to dynamic character movements.
#KlingAI #Kling4
🤝 Paid partnership · 🤖 Made with AI
For this test, I went with a "cyberpunk hacker robot working in front of a wall of monitors".
Prompt: "Cyberpunk hacker robot working in front of many monitors."
I like how the scene moves between the wider setup, close-ups of the robot’s mechanical hands, and all those code-filled screens. The different angles make the futuristic workspace feel more immersive, while the robot’s consistent look ties the whole sequence together.
Try it out 👀
kling.ai
For the second test, I took the same robot into a completely different setting: a futuristic city street.
Same black metallic armor, same glowing red eyes, but now it’s walking toward the camera before stopping to speak.
The prompt was: "The same robot walks through a rainy neon street at night, stops, looks into the camera, and says: Access granted."
This is what makes Omni Reference interesting to test, as you can take a character from one scene into another without reinventing its look from scratch. A tip: use a clear frame from your first generation as the reference, then describe the new environment and the character’s actions in sequence.
MiniMax has published MiniMax M3.1 Flash weights on HuggingFace.
> The model has 428B parameters total, 23B activated per token.
> MiniMax M3.1 Flash, also known as "Space Bunny" stealth model from OpenRouter, was announced on October 19 and is now available on MiniMax platform and APIs.
MiniMax M3.1 Flash weights are now public. 428B parameters total, 23B activated per token.
Agentic performance: A generational leap in agentic tasks, coding, and cybersecurity — 76.2 on BankerToolBench, 73.7% on DeepSWE v1.1, 61.7 on OSWorld V2, 29.6 on Agents' Last Exam, and 82.2 on CyberGym.
Efficient by design: MiniMax Sparse Attention and DSpark speculative decoding improve inference efficiency.
Hugging Face:
huggingface.co/MiniMaxAI/Min…
ModelScope:
modelscope.cn/models/MiniMax…
API: platform.minimax.io
M Plan: minimax.io/m-planMiniMax
MCode for desktop: agent.minimax.io/download
An upcoming Kimi K3.1 model release was officially teased!
Earlier, Kimi K3.1 has been spotted in testing on Arena and reportedly has a very good performance.
Soon 👀
Documented 🗞️
testingcatalog.com/kimi-k3-1…
DAILY AI BRIEF 🗞 — Oct 23
ANTHROPIC 🔥:
- Claude Models on Google Cloud (Anthropic-managed) is now generally available, bringing the full Claude Platform feature set with Google Cloud IAM and billing.
- Starting next week, Claude web, desktop, and mobile sessions will run in auto mode by default on Pro, Max, and Team plans, with an automated check blocking risky actions.
- Claude Duet will be optimized for Apple's foldable iPhone Duo.
OPENAI 🔥:
- The Decisions API is now generally available with GPT-6 Luna, offering near real-time classifications and decisions up to 10x faster at $0.10 per million input tokens.
SPACEXSI 🔥:
- Grok Bot can now use Calendar, Reminders, Contacts, and more apps on your phone, rolling out on iOS and Android.
- Grok Bot can now use your 1Password account to sign in with logins you approve, rolling out over the next few days.
GOOGLE 🔥:
- Google Cloud introduced Agent Compute Engine, a new VM type for running AI agents at scale, alongside expanded agent governance tools.
- Gemini can now reference your Google Maps searches and saved places alongside Google Calendar for tailored daily recommendations.
MINIMAX 🔥:
- MiniMax opened the weights of M3.1 Flash, a 428B-parameter MoE model with 23B active parameters, with day-0 support from Together, SGLang, vLLM partners, and NVIDIA NeMo.
MOONSHOT 🔥:
- Moonshot officially teased an upcoming Kimi K3.1 release after the model was spotted performing strongly in Arena testing.
MICROSOFT 🔥:
- Microsoft is rolling out its redesigned Copilot app to more users on web and mobile, with voice mode now appearing to require a paid plan.
PERPLEXITY 🔥:
- Perplexity added a Batch API to its Agent API supporting up to 50k requests per file, and its Decision Model can now run fully locally on a Mac.
CANVA 🔥:
- Canva launched Craft, its first in-house frontier image model built for design, available to all users and via API.
IDEOGRAM 🔥:
- Ideogram released Vector 1.0, a model that generates editable, layered SVGs and vectorizes images, available in Ideogram and via API.
1PASSWORD 🔥:
- 1Password announced an Agents platform with Meta, Anthropic, OpenAI, SpaceXSI, and others, so personal agents can work without seeing your passwords, with a waitlist open.
* Used Grok to compose this brief, cherry-picking the news and doing some post-editing.
DAILY AI BRIEF 🗞 — Oct 22
BREAKING 🔥:
- PHP 8.5.12, 8.4.27, 8.3.36 and 8.2.35 fix two critical unauthenticated remote code execution flaws dubbed php2shell (CVE-2026-109618 and CVE-2026-109619), which OpenAI's Patch the Planet initiative helped uncover and now offers a Codex skill to check affected sites.
- Microsoft Threat Intelligence reported that China state-aligned threat actor Storm-0940 used the agentic AI penetration testing framework ARTEX to automate attacks on hundreds of organizations worldwide.
- Cloudflare warned that private keys for eight widely trusted eMudhra intermediate certificates were compromised, letting attackers forge certificates for any domain.
- Node.js shipped security releases for its 22.x, 24.x and 26.x lines fixing multiple vulnerabilities, including high-severity issues.
- Cloudflare warned that CVE-2026-88771 lets attackers run commands on NetScaler Gateway before anyone signs in.
ANTHROPIC 🔥:
- Anthropic opened a waitlist for Claude Duet, a standalone proactive agent app built around Claude Opus 5.5 that can also shop via Stripe Link and Shop Pay, for Pro and Max users in the US.
- Claude Security now re-scans connected repositories as code changes land, in public beta on Claude Enterprise.
- Anthropic released API version 2026-10-21 with new SDKs and ant CLI 2.0, now licensed under Apache 2.0.
OPENAI 🔥:
- GPT-6 Astra, GPT-6.1 Sol and GPT-6 Luna plus Codex workflows are rolling out in private preview on Google Cloud's Model Garden.
- ChatGPT Ads expanded to 26 new countries, with more than 100,000 advertisers on the platform so far.
SPACEXSI 🔥:
- Grok Bots are rolling out to X Chat, letting you message your bots like any contact without leaving the X app.
- Grok Bot can now build reports, memos and dashboards as live pages you can edit together or have it keep up to date.
GOOGLE 🔥:
- Gemini is now available in Google Maps navigation on CarPlay for parking, traffic reports and route changes by voice.
META 🔥:
- Meta reported $64.6B in Q3 revenue, up 26%, with Zuckerberg saying more than 25% of Muse users have bought items through the agent.
NVIDIA 🔥:
- At GTC Berlin, NVIDIA introduced the OpenShell Reference Architecture for securing enterprise agents and DSX POD, cutting AI factory deployment at prepared sites to as little as 14 days.
- NVIDIA released Nemotron 3.5 Embed, Rerank and Parse open retrieval models on Hugging Face.
MISTRAL 🔥:
- Mistral partnered with NVIDIA on AI for industrial engineering, covering faster simulation, open physics-AI benchmarks and agentic evals for CAE and EDA.
KLING 🔥:
- The Kling 4.0 API is now officially open, with the model also landing on Runway, fal, Higgsfield and Adobe Firefly.
COHERE 🔥:
- Cohere launched Rerank 5 in Pro and Fast tiers.
RECURSIVE 🔥:
- Recursive released Flint, a 40B MoE open-weight local coding model under Apache 2.0 that it says is up to 2x faster than comparable models.
* Used Grok to compose this brief, cherry-picking the news and doing some post-editing.
"Daily AI Brief" also arrives in email format from Tuesday to Saturday.
OPT-IN below 👀 testingcatalog.com/#/portal/…
Anthropic ❤️ Google Cloud
Claude Platform features are now available on GCP!
> “With the Anthropic managed offering, Google Cloud customers now receive the full set of Claude API features, with new features and betas shipping the same day they go live on the native Claude Platform.”
This solution covers Claude Managed Agents, Code Execution, Computer Use, Agent Skills and Advisor Strategy.
Google is a Platform of Platforms now 👀
SPACEXSI 🔥: Grok Bot can now be connected to system mobile apps including Calendar, Reminders, Home and more!
This capability is rolling out on both iOS and Android.
So far, I have only seen this integration in Perplexity and Claude. This is quite a big unlock, at least up to a level that iOS and Android provide.
I keep using it more and more 👀
OPENAI 🔥: Decisions API is now in general availability and can be used via APIs and the OpenAI Platform.
> Decisions API is a Jev-like solution enabling you to build workflows based on extremely quick decisions.
> Predicate, Choice, and Score decisions are available.
It thinks that there is a 21% chance that the GPT-7 family will be released this year.
Looks right 👀
Kylon launched Annotations, a feature that lets users take a video they like and create a new version for their own product.
Kylon first analyzes the original video, breaks it into components, and then rebuilds it around the user's brand.
> Kylon agent can be tagged directly on X as @kylon_ai under any video post.
> Users can adjust pre-made prompts to start video generation on Kylon directly in order to see a full process breakdown.
> Teams can collaborate on Kylon by commenting on video components to steer production.
🤝 Paid partnership
Copying is easy. Making it yours is hard.
Meet Kylon Annotations: analyze, annotate and remix to add your team's taste together.
Drop a video below and @kylon_ai , we'll remix for you.
We tagged @kylon_ai under Quinn Leng's launch video and asked for a TestingCatalog version too!
Remaking it on the web requires a Kylon account, which starts with a 7-day trial. Kylon is also offering the first month for free to users who quote and repost its launch post.
Test it out 👀
kylon.io/fork
