After evaluating one of our upcoming models, Astra, we're treating it as our first "critical" model for cybersecurity under our Preparedness Framework.
This is a scenario we've planned for, and we're putting additional controls in place to ensure Astra's further development happens safely and securely.
We're working hard to make Astra broadly available, and get its advanced cyber capabilities into the hands of defenders.
https://t.co/9lMIvdeMMJ
DeepSeek V4 Flash from @deepseek_ai on ARC-AGI (Verified):
- ARC-AGI-2: 61.4%, $0.04/task
- ARC-AGI-1: 89.0%, $0.02/task
DeepSeek V4 Flash sets the new standard on the cost-to-performance Pareto frontier.
A ‘WARHAMMER 40,000’ animated series is in the works with Henry Cavill set to exec produce.
The series will be a spin-off of the Warhammer episode in ‘Secret Level’, with Blur Studios returning to animate.
(Source: https://t.co/qx5LnoZRit)
We tested DeepSeek V4 Flash against Kimi K3 and GLM 5.2 on 30 challenging agentic tasks.
DeepSeek was 2.5x faster than GLM and 1.4x faster than Kimi at a similar success rate, despite using the most tokens per task. 🧵🧵🧵
Yeah, this new Deepseek model is replacing grok 4.5 and glm. It may not one shot every time but it is really good at fixing itself, that’s good enough for me.
But then it will need to be connected to the Internet for that to happen and that just creates a whole new set of problems.
It should have a qr code scanner though.
I don't understand why I have to punch codes into that meter thingy. Why does the unit not automatically reflect on the meter like airtime does on phones?