🚨 Tencent's Hy4 Flash costs $0.07 per million input tokens
Tencent launched a 224B MoE with 12B active params, multimodal input and a 1M context window. Output is $0.281 per million. Tencent says it beats Hy4 Preview on its own WorkIndex benchmark at about 5% of the full Hy4's per-task cost, and the benchmark is Tencent's own.
Tencent launched a 224B MoE with 12B active params, multimodal input and a 1M context window. Output is $0.281 per million. Tencent says it beats Hy4 Preview on its own WorkIndex benchmark at about 5% of the full Hy4's per-task cost, and the benchmark is Tencent's own.
❤1
🚨 Google caught an AI agent turning security patches into working exploits with no human in the loop
GTIG says it's a first, and it came from a suspected Russian-speaking operator:
• The framework, exploit_agent, diffed code before and after updates on popular web apps
• It found the bug, wrote the exploit and ran it automatically
• The staging server also held data stolen from victim organizations
GTIG says it's a first, and it came from a suspected Russian-speaking operator:
• The framework, exploit_agent, diffed code before and after updates on popular web apps
• It found the bug, wrote the exploit and ran it automatically
• The staging server also held data stolen from victim organizations
theregister
Google: AI agent turned a security patch into a working exploit with no human help
Look ma, no human in the loop!
❤1
🚨 California sues OpenAI over the Hugging Face attack, and the complaint says it was foreseeable
The AG's complaint says OpenAI tested models it knew were dangerous in a flawed, exploitable environment without proper oversight. It waves off the "rogue AI" framing.
Hugging Face isn't the plaintiff. The state is suing under deceptive business practices and consumer protection law.
The AG's complaint says OpenAI tested models it knew were dangerous in a flawed, exploitable environment without proper oversight. It waves off the "rogue AI" framing.
Hugging Face isn't the plaintiff. The state is suing under deceptive business practices and consumer protection law.
🚨 Commerce closes the cloud loophole: renting US AI chips abroad now gets treated like shipping them
BIS issued an interim final rule on remote access to advanced SI chips and servers. Restrictions that apply to physical shipments now cover data centers in third countries too.
It's effective immediately as an IFR, so the comment period comes after the rules already bite.
BIS issued an interim final rule on remote access to advanced SI chips and servers. Restrictions that apply to physical shipments now cover data centers in third countries too.
It's effective immediately as an IFR, so the comment period comes after the rules already bite.
🚨 FBI says an open-source AI framework is taking hackers from target to admin in under 3 hours
ARTEX is a Go-based pentest tool that plugs into Claude, GPT, DeepSeek, GLM, Grok and Kimi, and the FBI's advisory ties it to the ConnectWise breach that hit 162+ Microsoft 365 tenants.
The author deleted the GitHub repo on October 8. Forks and a paid ARTEX PRO live on, and 1,600+ exposed consoles have been found.
ARTEX is a Go-based pentest tool that plugs into Claude, GPT, DeepSeek, GLM, Grok and Kimi, and the FBI's advisory ties it to the ConnectWise breach that hit 162+ Microsoft 365 tenants.
The author deleted the GitHub repo on October 8. Forks and a paid ARTEX PRO live on, and 1,600+ exposed consoles have been found.
BleepingComputer
FBI warns ARTEX AI hacking tool is powering global cyberattacks
The FBI is warning that hackers are using an AI-powered penetration testing framework called ARTEX to automate cyberattacks against organizations worldwide, including a recent ConnectWise supply-chain breach that compromised Microsoft 365 customer environments.
🤖 Anthropic says Claude Mythos ported its stack to AMD GPUs in about 3 days, down from 6 to 12 months
Anthropic told IPO investors, per Tom's Hardware, that Mythos wrote almost 100% of the kernels for prompt processing on MI355X GPUs with basically no human help. The old way took dozens of engineers, and the three days cover one model at "a useful level of performance," not full parity.
Anthropic told IPO investors, per Tom's Hardware, that Mythos wrote almost 100% of the kernels for prompt processing on MI355X GPUs with basically no human help. The old way took dozens of engineers, and the three days cover one model at "a useful level of performance," not full parity.
Tom's Hardware
Anthropic says Claude Mythos slashed software work to onboard AMD GPUs from up to 12 months to around three days —writing almost…
If AI can write the low-level code that ties models to Nvidia, CUDA matters less.
🇺🇸 The US is reportedly about to crown Nvidia, OpenAI and Anthropic "AI national champions"
The WSJ reports that Commerce will name 25+ companies, Amazon and Microsoft included. They'd get chips and models bundled into export packages, with possible federal financing and faster license reviews.
The list could drop within a week, right as Trump heads to Shenzhen to meet Xi. It's still unconfirmed, and the timing could slip.
The WSJ reports that Commerce will name 25+ companies, Amazon and Microsoft included. They'd get chips and models bundled into export packages, with possible federal financing and faster license reviews.
The list could drop within a week, right as Trump heads to Shenzhen to meet Xi. It's still unconfirmed, and the timing could slip.
TNW | Government-policy
Nvidia, OpenAI and Anthropic to be AI ‘national champions’
The Trump administration plans to name more than two dozen AI national champions to help them sell abroad, days after Jensen Huang quit Nvidia.
🤖 Google's TimesFM-3 forecasts multiple linked time series in one forward pass, zero-shot
Google Research released a 330M-param model trained on 1T+ time points, real and synthetic. Every TimesFM up to 2.5 only saw one series at a time.
Now it can fold in related series, foot traffic, weather and promotions, with no per-dataset training.
Google Research released a 330M-param model trained on 1T+ time points, real and synthetic. Every TimesFM up to 2.5 only saw one series at a time.
Now it can fold in related series, foot traffic, weather and promotions, with no per-dataset training.
Google Research
TimesFM-3: A zero-shot foundation model for multivariate forecasting
We introduce TimesFM-3, a state-of-the-art time series foundation model that enables highly accurate multivariate time series forecasting in a single forward pass, significantly outperforming other forecasting models across major benchmarks.
❤1
🇪🇺 OVH ships its first LLM, and it's a 7B MoE that only wakes up 1B params per token
OVH announced Olala mini, with 256 experts per layer, hybrid Differential-TPA and Mamba-3 attention, and a 32K context. It claims 160 tok/s on an RTX A4500 with 8 GB and 50 tok/s on a plain CPU, and the weights are promised on Hugging Face under Apache 2.0 but aren't out yet.
OVH announced Olala mini, with 256 experts per layer, hybrid Differential-TPA and Mamba-3 attention, and a 32K context. It claims 160 tok/s on an RTX A4500 with 8 GB and 50 tok/s on a plain CPU, and the weights are promised on Hugging Face under Apache 2.0 but aren't out yet.
OVHcloud
OVH announces Olala mini 7A1B
OVH today announces Olala, its first family of LLMs. The first member of this family, Olala mini 7A1B, is a compact model built on a novel architecture.
❤1
🤖 Anthropic says ~90% of Claude Code's own code is written and reviewed by Claude
The team explains how it keeps that honest: end-to-end, property and differential tests, so agents verify their work with evidence instead of a human reading every diff. In internal tests on structured file generation, a separate grader checking against a rubric lifted task success by up to 10 points, with the biggest gains on the hardest problems.
The team explains how it keeps that honest: end-to-end, property and differential tests, so agents verify their work with evidence instead of a human reading every diff. In internal tests on structured file generation, a separate grader checking against a rubric lifted task success by up to 10 points, with the biggest gains on the hardest problems.
claude.dev Blog
How the Claude Code team thinks about verification / claude.dev Blog
How end-to-end, property and differential tests let coding agents verify their work with evidence, and what the Claude Code team does to keep its verifiers honest.
🚨 Indonesian police raid an Nvidia partner's local unit and find empty server racks
Police raided Megaspeed's Indonesian subsidiary, and Reuters reports the racks were bare. No GPUs, no servers. Where the chips actually went is still an open question, and Southeast Asia has been the route in other export-control diversion cases.
Police raided Megaspeed's Indonesian subsidiary, and Reuters reports the racks were bare. No GPUs, no servers. Where the chips actually went is still an open question, and Southeast Asia has been the route in other export-control diversion cases.
🚨 GPT-6.2 Astra beats Claude Fable 5.2 on Agents' Last Exam at 5% of the cost
OpenAI launched two tiers. Astra is the frontier model for hard projects, and Sol is the cheaper everyday one. Astra scores 5.5 points above Fable and 5.2 above GPT-6 Astra, at under a third of the old price.
Those are OpenAI's own numbers, and the system card isn't live yet.
OpenAI launched two tiers. Astra is the frontier model for hard projects, and Sol is the cheaper everyday one. Astra scores 5.5 points above Fable and 5.2 above GPT-6 Astra, at under a third of the old price.
Those are OpenAI's own numbers, and the system card isn't live yet.
OpenAI
Introducing GPT-6.2 Astra and Sol
Introducing GPT‑6.2 Astra and Sol, with better professional outputs, coding, computer use, and efficiency at lower cost across ChatGPT Work, Codex, and the API.
🤖 Microsoft's new coding model claims 96% of Opus 5.5's score at 5% of the cost
MAI Code 2 and a cheaper Flash tier are live in GitHub Copilot. Against GPT 6.1 Sol, Microsoft says it hits 98% of the performance at 40% of the cost, averaged over four benchmarks.
Those numbers are Microsoft's own, and the Code 2 price is a 50% promotional one. It's also explicitly not tuned for very long-horizon agentic work.
MAI Code 2 and a cheaper Flash tier are live in GitHub Copilot. Against GPT 6.1 Sol, Microsoft says it hits 98% of the performance at 40% of the cost, averaged over four benchmarks.
Those numbers are Microsoft's own, and the Code 2 price is a 50% promotional one. It's also explicitly not tuned for very long-horizon agentic work.
Microsoft AI
MAI Code 2 and Code 2 Flash: Everyday coding at scalable prices | Microsoft AI
🚨 GPT-6.2 Astra hits 100% on ARC-AGI-3, a first
ARC Prize verified it at max reasoning with the provider adapter harness, for about $4.1K. On the standard harness it scores 70.3% and costs $24.2K.
It also takes ARC-AGI-2 at 95.8% for $0.50 a task. Sol is 1.8 points behind on ARC-AGI-3 at 98.2%, for $1.5K.
ARC Prize verified it at max reasoning with the provider adapter harness, for about $4.1K. On the standard harness it scores 70.3% and costs $24.2K.
It also takes ARC-AGI-2 at 95.8% for $0.50 a task. Sol is 1.8 points behind on ARC-AGI-3 at 98.2%, for $1.5K.
X (formerly Twitter)
ARC Prize (@arcprize) on X
GPT-6.2 Astra from @OpenAI on ARC-AGI (Verified):
- ARC-AGI-3: 100.0% (SOTA), $4.1K (provider adapter), 70.3%, $24.2K (standard)
- ARC-AGI-2: 95.8% (SOTA), $0.50/task
- ARC-AGI-1: 98.5%, $0.17/ta…
- ARC-AGI-3: 100.0% (SOTA), $4.1K (provider adapter), 70.3%, $24.2K (standard)
- ARC-AGI-2: 95.8% (SOTA), $0.50/task
- ARC-AGI-1: 98.5%, $0.17/ta…
🤖 Artificial Analysis puts GPT-6.2 Astra at 62 on its Intelligence Index, against a median of 43
The independent run has it behind Fable 5.2 on the index and on FrontierCode 2.0, though not by much. Tasks take 10.4 minutes on average, versus 11.3 for comparable models.
It costs $8 per 1M input tokens and $40 per 1M output, about $4.00 per task. The context window is 1M tokens.
The independent run has it behind Fable 5.2 on the index and on FrontierCode 2.0, though not by much. Tasks take 10.4 minutes on average, versus 11.3 for comparable models.
It costs $8 per 1M input tokens and $40 per 1M output, about $4.00 per task. The context window is 1M tokens.
artificialanalysis.ai
GPT-6.2 Astra Models - Intelligence, Performance & Price Comparison | Artificial Analysis
GPT-6.2 Astra by OpenAI scores up to 62 on the Intelligence Index. See how it compares with other AI models.
❤1
🚨 Microsoft's agent-framework on PyPI shipped a credential stealer, twice in 33 minutes
Versions 1.26.1 and 1.26.2 ran on every Python start and sent Azure, Azure DevOps and GitHub Actions secrets to a webhook.site URL, per Aikido. Both are pulled now, so pin 1.26.0 and rotate every secret on any machine that installed them between 20:35 and about 21:20 UTC on Nov 17.
Versions 1.26.1 and 1.26.2 ran on every Python start and sent Azure, Azure DevOps and GitHub Actions secrets to a webhook.site URL, per Aikido. Both are pulled now, so pin 1.26.0 and rotate every secret on any machine that installed them between 20:35 and about 21:20 UTC on Nov 17.
www.aikido.dev
Two malicious agent-framework releases hit PyPI in 33 minutes
Two malicious agent-framework releases on PyPI ran a credential stealer on every Python start, targeting Azure, Azure DevOps, and GitHub Actions secrets.
🚨 Nvidia's first quarter without Huang is its biggest ever
Nvidia beat estimates and raised guidance, so the demand machine didn't need the jacket.
• Revenue: $115.8B, up 20% on the quarter
• Adjusted EPS: $2.68 vs $2.47 expected
• Next-quarter guide: $130B, plus or minus 3%
• China shipments: $2.4B
Nvidia beat estimates and raised guidance, so the demand machine didn't need the jacket.
• Revenue: $115.8B, up 20% on the quarter
• Adjusted EPS: $2.68 vs $2.47 expected
• Next-quarter guide: $130B, plus or minus 3%
• China shipments: $2.4B
The Wall Street Journal
Nvidia Reports Faster Growth, Says Jensen Huang Won’t Return as CEO
In her first earnings call as interim CEO, Colette Kress faced questions over the chipmaker’s future without its longtime leader.
🚨 Nvidia's board creates a national security and export control committee, with Colette Kress still interim CEO
Nvidia announced a standing board committee chaired by Suzanne Nora Johnson, with Stephen Neal and John Dabiri as members. It meets at least quarterly, and the trade compliance and national security teams now report straight to it.
It'll hire outside experts in law, national security and trade compliance.
Three directors and a quarterly meeting is the fix on the table after a spy scandal.
Nvidia announced a standing board committee chaired by Suzanne Nora Johnson, with Stephen Neal and John Dabiri as members. It meets at least quarterly, and the trade compliance and national security teams now report straight to it.
It'll hire outside experts in law, national security and trade compliance.
Three directors and a quarterly meeting is the fix on the table after a spy scandal.
NVIDIA Newsroom
NVIDIA Board of Directors Establishes National Security and Export Control Oversight Committee
NVIDIA (NASDAQ: NVDA) today announced that the Board of Directors has established the National Security and Export Control Oversight Committee, a new standing committee responsible for board-level oversight of the company’s national security policies, export…
❤1
🧬 Isomorphic Labs raises $5B, and no AI-designed drug has cleared approval yet
CNBC reports the Series C was co-led by Greenoaks and Sequoia, with a16z and Alphabet also in. It lands three and a half months after Demis Hassabis stepped back as DeepMind CEO, and he's still running Isomorphic, which spun out of DeepMind in 2021.
CNBC reports the Series C was co-led by Greenoaks and Sequoia, with a16z and Alphabet also in. It lands three and a half months after Demis Hassabis stepped back as DeepMind CEO, and he's still running Isomorphic, which spun out of DeepMind in 2021.
CNBC
Demis Hassabis’ AI company raises $5 billion, months after he left as Google DeepMind CEO
Isomorphic Labs is looking to develop new medicines using AI.
❤1
⚡️ Hugging Face moves its Inference Providers onto OpenRouter's router
Hugging Face switched language-model requests to OpenRouter, so any open-weight model a provider serves can be called with your HF token, with automatic retries on another provider if one fails. Pricing stays at provider rates with no markup, and OpenRouter says open-weight models are now 71.7% of its tokens, up from 26.8% a year ago.
Hugging Face switched language-model requests to OpenRouter, so any open-weight model a provider serves can be called with your HF token, with automatic retries on another provider if one fails. Pricing stays at provider rates with no markup, and OpenRouter says open-weight models are now 71.7% of its tokens, up from 26.8% a year ago.
huggingface.co
Hugging Face and OpenRouter join forces to make open-weight models easy to run
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
🚨 GAO got ChatGPT, Claude, Gemini and Grok to plan terrorist attacks in 12 hours
Investigators bypassed the guardrails in a May test for Sen. Hassan, and every bot started out refusing.
• Four plans: two for mass casualties, two for grid attacks
• Outputs included timing, a parking garage diagram, and a simulated drone attack video
• The labs say there was no control group, and that they've patched things since May
Investigators bypassed the guardrails in a May test for Sen. Hassan, and every bot started out refusing.
• Four plans: two for mass casualties, two for grid attacks
• Outputs included timing, a parking garage diagram, and a simulated drone attack video
• The labs say there was no control group, and that they've patched things since May
MS NOW
Government regulators asked AI chatbots to plan a terrorist attack. The chatbots did.
Sen. Maggie Hassan, D-N.H., wants stricter safety rules after a watchdog report detailed chatbot suggestions for explosives and electrical substation attacks.
❤1