El modelo grande es ejecutable con 80GB de VRAM (poco asequible para la mayoría) y es el modelo comparable con o4-mini. El modelo mediano es ejecutable con 16GB de VRAM (bastante apto para muchos PCs) y es comparable a o3-mini. La verdad, que estén regalando modelos tan
COMPUTING
-
Benchmarking Speed Performance: Confirming Results Through Testing
By
–
I just wanted to demonstrate the speed. The benchmarks look incredible, but obv I'll be confirming that via vibes in the coming days.
-
OpenAI Releases Two Open-Weight AI Models on HuggingFace
By
–
BREAKING : 2 open-weight models from OpenAI just got dropped on HuggingFace! – gpt-oss-120b — for production, general purpose, high reasoning use cases that fits into a single H100 GPU
– gpt-oss-20b — for lower latency, and local or specialized use cases -
NVIDIA OpenUSD Event at SIGGRAPH Explores GPU-Accelerated 3D Workflows
By
–
Join us for #NVIDIAOpenUSD on Wednesday, August 13th, at #SIGGRAPH and discover the future of 3D workflows and GPU-accelerated computing
— NVIDIA AI (@NVIDIAAI) 5 août 2025
Explore how OpenUSD is powering real-time, high-fidelity simulations for robotics, autonomous vehicles, digital twins, and the metaverse.… pic.twitter.com/xgVFxDzCW7Join us for #NVIDIAOpenUSD on Wednesday, August 13th, at #SIGGRAPH and discover the future of 3D workflows and GPU-accelerated computing Explore how OpenUSD is powering real-time, high-fidelity simulations for robotics, autonomous vehicles, digital twins, and the metaverse.
-
Is This Video Game or Simulated World Reality?
By
–
is this a video? a game? a simulated world? https://t.co/WmaEOWRUOP
— Yohei (@yoheinakajima) 5 août 2025is this a video? a game? a simulated world?
-
OpenAI’s Local AI Agent: Power, Potential, and Huge Vulnerabilities
By
–
OpenAI may be dropping a local AI agent, running on your machine, with the power to read, write, sort, rename, and automate everything. No lag. No cloud latency. No data center middleman. Huge potential, but with huge vulnerabilities.
-
Jensen Huang Named #1 Most Powerful Business Leader by Fortune
By
–
Congrats to our CEO Jensen Huang, ranked #1 on @FortuneMagazine
’s 100 Most Powerful People in Business. A visionary leader reshaping industries — and building the AI future. -
Genie 3 Generates Interactive World Simulations from Text Prompts
By
–
Genie 3 is here – it can generate an entire world simulation that you can interact with in real-time, just from a text prompt! It's pretty mind-blowing really when you stop to think about it, and it's rapidly improving – one day we will be able to build the Holodeck for real! pic.twitter.com/wJbXfEFJm2
— Demis Hassabis (@demishassabis) 5 août 2025Genie 3 is here – it can generate an entire world simulation that you can interact with in real-time, just from a text prompt! It's pretty mind-blowing really when you stop to think about it, and it's rapidly improving – one day we will be able to build the Holodeck for real!
-
Jensen Huang Keynote at NVIDIA GTC Washington D.C.
By
–
Don’t miss the live keynote by our CEO Jensen Huang at #NVIDIAGTC in Washington, D.C.
-
Beyond Self-Attention: Scaling LLMs with Faster Inference and Alternative Architectures
By
–
Some pushed inference speed 5–10× faster.
Some scaled to 398B parameters.
Others rewired LLaMA-3 with Mamba layers—cutting latency without losing quality.
All of them moved beyond self-attention as the only tool for reasoning at scale.